What Is the Prometheus Certification?

The Prometheus certification is the Prometheus Certified Associate, exam code PCA, run by the Linux Foundation with the Cloud Native Computing Foundation. It tests foundational observability knowledge and practical Prometheus skills: scraping targets, writing PromQL, instrumenting code and configuring alerts. It is 60 questions in 90 minutes, 75 percent to pass, and $250 for the exam only.

The purchase includes one retake and twelve months of exam eligibility from the date you buy. Two bundles are sold alongside the exam: the LFS241 course plus the exam for $299, and the exam with the THRIVE-ONE annual subscription for $495. If you only want the credential and already know Prometheus, the exam alone is the sensible buy.

What the credential actually proves

The Linux Foundation describes PCA as a pre-professional certification. It confirms that you can build or scrape observability data in an application stack, whether or not that stack is cloud native, and that you can use the resulting data to troubleshoot and to feed other systems. It does not claim you can run Prometheus at very large scale, and it does not cover long-term storage projects or managed services.

Who Is the Prometheus Certified Associate Written For?

The Prometheus Certified Associate is written for engineers and application developers with a particular interest in observability and monitoring. The Linux Foundation rates it beginner level and sets no prerequisites, though it names KCNA, CKA or CKAD holders, and people who have finished Prometheus training or a cloud engineer bootcamp, as the ideal candidates.

In practice that covers three groups. Site reliability and platform engineers who already run Prometheus and want proof of the fundamentals. Application developers who are asked to expose metrics from their own services and need to understand what happens to those numbers after scraping. And data engineers who treat metrics as one more time series feed, because Prometheus is, underneath, a labelled time series database with its own query language.

“Like the certifications before it, the PCA gives developers the ability to officially demonstrate their skills in understanding Prometheus and modern observability systems.”

The exam went generally available in September 2022 after several months of beta testing, as the CNCF launch announcement records. It has kept the same five domain structure since, which makes the published curriculum a reliable study map.

Who should probably wait

Someone who has never seen a metrics endpoint will find the paper hard despite the beginner label. The questions assume you have looked at real scrape output, written at least a few queries and read an alerting rule. A week of hands-on time with a local Prometheus server before booking is the cheapest way to close that gap, and working through a PCA practice exam afterwards shows which domains still cost you marks.

How Is the PCA Exam Weighted?

The PCA exam is weighted across five domains. PromQL is the largest at 28 percent, Prometheus Fundamentals carries 20 percent, Observability Concepts and Alerting and Dashboarding take 18 percent each, and Instrumentation and Exporters is the smallest at 16 percent. The Linux Foundation publishes these weights alongside the objective list, so preparation can follow them exactly.

A 75 percent pass mark on 60 questions means 45 correct answers, so you can afford to lose 15. That is roughly the size of the whole Instrumentation and Exporters domain, which is a useful way to picture the margin: there is room for one weak area, not two. The Linux Foundation does not publish a per domain question count, so the percentages are the only guide to how many questions each area produces. The same weights and objective list appear on the official PCA curriculum, which is worth rechecking before you book.

Why Does PromQL Carry 28 Percent of the Exam?

PromQL carries 28 percent of PCA because it is how every other part of Prometheus gets used. Dashboards are PromQL expressions, alerting rules are PromQL expressions with a threshold and a duration, and troubleshooting is PromQL typed into the expression browser. A candidate who cannot read a query cannot really use the data the other domains describe collecting.

The seven objectives are selecting data, rates and derivatives, aggregating over time, aggregating over dimensions, binary operators, histograms and timestamp metrics. They build on each other, so study them in that order rather than jumping straight to histograms.

Selecting data and the instant versus range distinction

Every query starts with a selector: a metric name and a set of label matchers, which can be equality, inequality or regular expression matches. The first concept that trips people is the difference between an instant vector, one sample per series at a moment, and a range vector, a window of samples per series written with a duration in square brackets. Many functions accept only one of the two, and a surprising share of wrong answers comes from passing the wrong kind.

Rates, counters and why rate() exists

Counters only go up and reset to zero when a process restarts, so their raw value is almost never what you want. The rate() function turns a counter range into a per second average and handles resets for you, irate() uses only the last two samples, and increase() expresses the same change over the whole window. Know which one suits a dashboard, which suits an alert, and why rate() should be taken before sum(), not after. The official PromQL querying basics page is the right depth for all of this.

Aggregation, operators and histograms

Aggregating over dimensions means operators such as sum, avg, max and count with by or without clauses that decide which labels survive. Aggregating over time means the over_time family, such as avg_over_time and max_over_time, applied to a range. Binary operators add arithmetic, comparison and set operations between vectors, together with the label matching rules that decide which series pair up. Histograms close the domain: histogram_quantile() estimates a quantile from cumulative bucket counters, and the exam expects you to know that rate() is applied to the buckets first.

What Do Prometheus Fundamentals and Instrumentation Test?

Prometheus Fundamentals, at 20 percent, tests the architecture, scrape configuration, the data model and labels, the exposition format and the system’s known limitations. Instrumentation and Exporters, at 16 percent, tests how metrics get into Prometheus in the first place: client libraries, instrumenting your own code, exporters for third party systems, and naming metrics well.

Architecture and scraping

Expect questions on the parts of a Prometheus deployment: the server that scrapes and stores, targets that expose an HTTP metrics endpoint, service discovery that finds those targets, Alertmanager that handles notifications, and the Pushgateway for short-lived batch jobs that cannot be scraped. Scrape configuration covers jobs, intervals, timeouts and relabelling, and you should be able to read a scrape_configs block and say what it will collect.

Data model, labels and limitations

A time series is identified by its metric name plus its full set of label pairs, so every distinct label combination is a separate series. That is why high cardinality labels, such as user IDs or request paths with unbounded values, are the classic way to overload a server. The limitations objective also covers what Prometheus is not designed for: it favours reliability of the monitoring system over perfect accuracy, keeps local storage on a single node, and is not a store for logs, events or exact billing data.

Instrumentation and the four metric types

Client libraries give you four metric types: counters, gauges, histograms and summaries. The exam asks when each fits, for example a counter for requests served, a gauge for queue depth, and a histogram rather than a summary when you need to aggregate quantiles across instances. Exporters translate metrics from systems you cannot instrument directly, such as the node exporter for host metrics. Naming follows the project’s conventions: a single unit in base units, a suffix such as _total for counters, and labels for dimensions rather than metric names.

How Are Observability Concepts and Alerting Examined?

Observability Concepts and Alerting and Dashboarding each carry 18 percent of PCA. The first tests the vocabulary around Prometheus: metrics versus logs, events and traces, push versus pull collection, service discovery and service level indicators. The second tests turning data into action: dashboard basics, writing alerting rules, and how Alertmanager routes, groups and silences notifications.

The concepts domain is broader than Prometheus

This is the one domain that is not purely about the tool. You need to explain why Prometheus pulls rather than waits for pushes, what tracing and spans record that metrics cannot, and how an SLI, an SLO and an SLA relate. The definitions in the Google SRE book chapter are the standard reference most exam questions on this objective lean on: an indicator is the measured quantity, an objective is the target for it, and an agreement is the contract with consequences attached.

Alerting rules and Alertmanager

The most common confusion here is about which component does what. The Prometheus server evaluates alerting rules, which are PromQL expressions with an optional for duration, and fires alerts when a condition holds long enough. Alertmanager never evaluates a query. It receives fired alerts, groups related ones, applies inhibition and silences, and routes the result to receivers such as email, chat or paging tools. The alerting basics objective adds judgement: alert on symptoms users feel rather than every cause, and make every page actionable.

Dashboarding basics

Dashboarding is the lightest objective in the domain. Expect questions on choosing a sensible query for a panel, rate windows that suit the scrape interval, and the difference between a dashboard for exploration and one for an on call view. The blueprint names no specific dashboard product, so concentrate on the queries behind panels rather than any tool’s menus. Every objective across all five domains is listed as a single checklist in our Prometheus Associate topic notes.

How Should You Prepare for the Prometheus Certified Associate?

Prepare for the Prometheus Certified Associate by following the weighting and doing everything against a running server, even though the exam is multiple choice. Three to five weeks is realistic for someone who already uses Prometheus at work. The sequence below spends the most time on PromQL and the fundamentals, which together carry 48 percent of the marks.

1. Run a local Prometheus server and a node exporter, then read the raw output of the metrics endpoint until the exposition format and the labels in it feel familiar.
2. Write the scrape configuration yourself, add a second job with service discovery or a static target, and break it on purpose so the targets page shows you the failure.
3. Work through PromQL in the order of the blueprint, from selectors to rates, aggregation, binary operators and histograms, predicting each result before you run the query.
4. Instrument a small application with a client library, exposing one counter, one gauge and one histogram with sensible names and units.
5. Write two alerting rules with a for duration, send them to a local Alertmanager, and practise grouping and a silence.
6. Finish with timed practice at 90 seconds per question, logging every wrong answer against the domain it came from so the remaining hours go to the weakest area.

“The market demand by both engineers and employers for a reliable baseline of Prometheus knowledge is proof positive that Prometheus and its ecosystem have crossed the chasm.”

Where to study from

The official Prometheus documentation is the primary source and covers every objective. LFS241 is the Linux Foundation’s own course for the exam if you prefer a guided path, and the bundle price is only $49 more than the exam on its own. Keep the curriculum open while studying: if a topic is not on it, it is not on the exam.

Where PCA Leads After You Pass

PCA is an associate credential, so it sits at the start of a path rather than the end. The natural next steps are either broader observability, through OpenTelemetry, or deeper Kubernetes skills, through the Linux Foundation’s administrator and developer exams. Which one fits depends on whether your work leans toward telemetry pipelines or toward running the platform itself.

For engineers whose job is collecting telemetry across services, OpenTelemetry is the obvious companion. It covers traces, metrics and logs with one set of APIs and a collector, and it is where the tracing and spans objective of PCA leads. Our OpenTelemetry associate guide covers that exam’s topics.

For platform engineers, Prometheus knowledge pays off most on Kubernetes, where the Prometheus Operator and service discovery are standard. The certificate itself stays valid for two years, so plan the next exam inside that window if you want the credentials to read as a continuous path.

Frequently Asked Questions

How many questions are on the Prometheus certification exam?

The Prometheus Certified Associate exam has 60 multiple-choice questions with a 90 minute limit. That works out to 90 seconds a question, so pace matters less than knowing PromQL well.

What is the passing score for PCA?

The passing score is 75 percent. On 60 questions that means 45 correct answers, leaving room to miss 15.

How much does the Prometheus certification cost?

The exam alone is $250 USD. The Linux Foundation also sells it with the LFS241 course for $299 and with the THRIVE-ONE annual subscription for $495. Each purchase includes one retake.

Are there prerequisites for the Prometheus Certified Associate?

No. Anyone can register. The Linux Foundation rates it beginner level and suggests that candidates holding KCNA, CKA or CKAD, or who have done Prometheus training, are a good fit.

Is the PCA exam hands-on?

No. It is an online, proctored, multiple-choice exam. Practising on a real server still helps, because many questions describe output you would only recognise from using one.

Which PCA domain carries the most weight?

PromQL, at 28 percent. Prometheus Fundamentals follows at 20 percent, Observability Concepts and Alerting and Dashboarding take 18 percent each, and Instrumentation and Exporters carries 16 percent.

How long is the Prometheus certification valid?

The certification is valid for two years. You have twelve months from purchase to sit the exam, including the retake.

Does the PCA exam cover Grafana?

The curriculum names dashboarding basics but no dashboard product. Expect questions about the queries behind panels rather than about any particular tool’s interface.

Does Alertmanager evaluate alerting rules?

No. The Prometheus server evaluates alerting rules and fires alerts. Alertmanager receives them, groups them, applies silences and inhibition, and routes notifications to receivers.

How long should I study for PCA?

Three to five weeks suits someone who already uses Prometheus. Newcomers should add a week or two of hands-on time with a local server before starting exam preparation.

Conclusion

The Prometheus certification rewards people who can read and write queries, not just people who have kept a monitoring stack running. PromQL is the largest of the five domains at 28 percent, and alerting rules and dashboards both depend on it, so its real share of the exam is larger still.

Build your preparation the same way: a local server, scrape configuration written by hand, PromQL in blueprint order, one instrumented application and two alerting rules routed through Alertmanager. With 60 questions, 90 minutes and a 75 percent pass mark, there is room for one weak domain. Make sure it is not the query language, then book the exam.

Is There a Cloudera Data Scientist Certification?

No. Cloudera’s certification catalogue has no data scientist exam. The Cloudera Machine Learning Engineer certification, CDP-6001, is the credential Cloudera offers to anyone who builds, tunes and deploys models on its platform, and it is the exam a data scientist on Cloudera would sit. The catalogue lists it as Cloudera Machine Learning Engineer; the exam guide calls it CDP Machine Learning Engineer; AnalyticsExam calls it Cloudera CDP Machine Learning Engineer.

The three names describe one exam with one code. Cloudera dropped the CDP prefix from its catalogue titles while keeping it in the codes, so CDP-6001 appears beside Cloudera Generalist, Cloudera Data Engineer and the two administrator exams. The guide’s audience statement is the clearest description of who the exam is for: a machine learning engineer who designs and develops models using MLOps and Cloudera Machine Learning, with a strong grasp of data modeling and data science concepts, deploying and tuning models, Spark, Spark MLlib, algorithms and general machine learning.

Read that list again and notice the order. Platform and MLOps come first, Spark second, algorithms last. The weightings in the next sections follow the same order, which is why a candidate with a strong statistics background and no Cloudera time is the one most likely to be surprised.

What Does the Cloudera Machine Learning Engineer Certification Cost and Cover?

CDP-6001 costs $330, runs 90 minutes and has 45 questions with a 60 percent pass mark. It is delivered online only, proctored live through Questionmark, with no reference material of any kind allowed. The two Cloudera courses that cover the content are optional. A pass is valid for two years, and the certification is for Cloudera Machine Learning Engineer, under whichever of its three titles you meet.

Two of those rows deserve a second look. The 60 percent pass mark is published by Cloudera on the CDP-6001 exam guide itself, which is not true of every Cloudera exam, so there is no guessing involved: 27 correct answers out of 45 is the line. And the price is the only figure that does not appear on the guide; it is on the Cloudera Education store page where the exam is bought, and a retake costs the full amount again.

What the question style looks like

The questions are scenario items set inside a named company, with four or five options and the occasional choose-two. The scenarios are short and the distractors are specific: one option names the right Cloudera feature for the wrong purpose, another names the right purpose for the wrong feature. That shape rewards candidates who have clicked through a workbench, not candidates who have memorised a feature list.

How Are the Five CDP-6001 Domains Weighted?

CDP-6001 has five domains. Cloudera Machine Learning carries 31 percent, Spark MLlib 22, Spark 18, Deploying a Machine Learning Model 18, and Deep Learning and General Machine Learning 11. The platform domain alone is nearly a third of the exam, Spark and MLlib together are 40 percent, and the general machine learning theory that most candidates already know is the smallest share.

Rounded to 45 questions, the weights translate to roughly 14 platform items, 10 on MLlib, 8 on Spark, 8 on deployment and 5 on general theory. A candidate who knows the platform and Spark cold can miss every theory question and still clear 60 percent; a candidate who knows only theory cannot pass. The published CDP-6001 sample questions show the mix in practice: items on runtimes, workspace quotas and data visualizations outnumber the ones on validation sets and algorithms.

Reading the platform domain

Seven topics sit under Cloudera Machine Learning, and they are the seven things a user does in the workbench in order. A workspace is provisioned by an administrator and holds projects and a shared compute quota. A project holds code, sessions, jobs and collaborators. An experiment records training runs and their metrics. Accelerators for ML Projects, called AMPs in the product, are complete reference projects built by Cloudera’s research arm, Fast Forward Labs, that install from a catalogue. Data Visualizations build shareable dashboards without leaving the project. Runtimes bundle an editor, a kernel and a library edition into one selectable image. GPUs are requested per session or job through a resource profile. Expect a question or two on each.

Why Does the Syllabus Say Cloudera Machine Learning When the Docs Say Cloudera AI?

Because the product was renamed after the exam guide was written. CDP-6001’s largest domain is titled Cloudera Machine Learning, and the sample questions abbreviate it to CML. Cloudera’s current documentation and product pages call the same service Cloudera AI, its workbench Cloudera AI Workbench, and the documentation URL path still reads machine-learning. Nothing in the domain changed; only the labels did.

This matters for one practical reason: the search terms in your head will not match the page titles in front of you. The Cloudera AI documentation is the right reference for every topic in the 31 percent domain, and it opens with the line that Cloudera AI is Cloudera’s cloud-native machine learning platform. Within it, the names map as follows.

The last row is the one to be careful with. The exam guide’s deployment domain describes models served from inside the workbench as REST APIs with autoscaling and metrics, and the documentation still describes exactly that feature, but it now labels it the previous generation and points new production work to the AI Registry and the AI Inference service. The exam has not been rewritten around those newer services. Study the workbench model feature the guide names, know that the documentation calls it legacy, and do not assume a question about the registry unless the guide is updated.

What Does the Spark Half of CDP-6001 Expect?

Spark and Spark MLlib together make up 40 percent of the Cloudera Machine Learning Engineer certification. The Spark domain covers DataFrames, file types and window functions; the MLlib domain covers pipelines, fitting and evaluating models, and model selection and tuning. These are PySpark coding questions in scenario form: which operation, which storage level, which file format, which evaluator, which parameter grid.

Spark: 18 percent

The three listed topics are narrower than they look. DataFrames questions test transformations that candidates get wrong under pressure: a groupBy with a pivot to turn months into columns, the difference between cache and persist with an explicit storage level, when a join should be broadcast. File types means knowing what Parquet, ORC, Avro, CSV and JSON are each good for, down to which columnar format carries lightweight indexes that let a reader skip row groups. Window functions means ranking, lag and lead, and running totals partitioned by a key, the operations that turn an event log into model features.

Spark MLlib: 22 percent

This is the second largest domain and the one most purely about code. The Apache Spark MLlib guide organises the DataFrame-based API around exactly the three listed topics: pipelines made of transformers and estimators, fitting and evaluating with the classification, regression and clustering evaluators, and model selection with cross-validation or train-validation split over a parameter grid. Expect questions on which stage goes where in a pipeline, what an evaluator’s metric name is, how a ParamGridBuilder expands, and why a validation set is held out separately from the test set.

The general machine learning domain, at 11 percent, is where the theory lives: supervised against unsupervised learning, the algorithm families, what overfitting looks like in a learning curve. If you have built models anywhere, this domain is already covered. Spend the time on the other 89 percent.

How Is Model Deployment Tested on the Cloudera ML Engineer Exam?

Deploying a Machine Learning Model is 18 percent of CDP-6001 and lists four topics: Applications/API, Autoscaling and Performance, Model Metrics and Monitoring, and ML Flow. In the product, deploying a model means selecting a Python or R function in a project file and having the workbench snapshot the code, package an immutable artifact, expose a REST endpoint and run a chosen number of load-balanced replicas.

That workflow is what the questions are built on. Applications and API items ask what a model endpoint accepts and returns, and how an Application differs from a model: an Application hosts a long-running web app such as a Flask or Streamlit front end inside the project, while a model serves predictions. Autoscaling and performance items ask about replica counts and resource profiles. Metrics and monitoring items draw the line the sample questions draw explicitly: comparing many training runs before deployment is experimentation, tracking one live model version is monitoring. ML Flow refers to the experiment tracking that records parameters and metrics for each run.

“Working with complex data environments every day, we know that moving from AI experimentation to production requires both the freedom to innovate and confidence that data remains governed and under our control”

That move from experimentation to production is the arc of the whole exam: workspace, project, experiment, deployed model, monitored endpoint. A candidate who can describe each hand-off, including which screen it happens on, has covered the platform domain and the deployment domain at once, which is 49 percent of the paper.

Who Should Take CDP-6001, and What Does the Role Pay?

CDP-6001 is written for machine learning engineers and data scientists who already work, or are about to work, on Cloudera. It is not an entry-level credential: the guide expects proficiency in designing and developing models with MLOps, plus Spark and MLlib expertise. Someone fluent in scikit-learn on a laptop but new to Spark and to the Cloudera workbench should budget most of their preparation for the platform and Spark domains.

On pay, PayScale puts the average base salary of a machine learning engineer in the United States at $126,315, with a range from $88,000 to $171,000 across 944 salary profiles updated in October 2026. No source attributes a premium to this credential specifically, and this article will not invent one. What the certification does is document, to an employer running Cloudera, that you can work inside its workbench rather than only around it.

Is the Cloudera certification worth it?

Worth it depends on one question: does your employer, or your next one, run Cloudera? If yes, CDP-6001 is the only machine learning credential on the vendor’s own list, and its platform domain is specific enough that a generic ML certification does not cover it. If no, a cloud vendor’s ML engineer credential will match more job listings. The three CDP-6001 pages BigDataRise already keeps, a syllabus-and-questions page, a Cloudera certification hub and this article, exist because the first answer is more common among its readers than it looks; the CDP-6001 exam topics page is the place to check the domain list again once you have read the guide.

How Should You Prepare for the Cloudera Machine Learning Engineer Exam?

Prepare for CDP-6001 in the order of its weightings: the workbench first, Spark and MLlib second, deployment third, theory last. Six to eight weeks is realistic for someone who already writes PySpark and has a Cloudera environment to practise in; longer without the environment, because 49 percent of the exam is about screens and features you need to have used.

1. Read the Cloudera exam guide once and note the two names for every platform feature, old and new, so the documentation search terms match.
2. Work through the Cloudera AI documentation for the seven platform topics in the order the syllabus lists them: workspaces, projects, experiments, AMPs, data visualizations, runtimes and GPUs.
3. Build one project end to end in a workbench: launch a session on a chosen runtime, log an experiment, deploy the model as a REST endpoint and read its metrics.
4. Rebuild your feature code in PySpark DataFrames, covering pivots, window functions, persist levels and the Parquet, ORC and Avro trade-offs.
5. Train one MLlib pipeline with a CrossValidator over a parameter grid and an evaluator, and explain each stage without notes.
6. Sit the sample questions under a 90 minute clock and keep a list of every feature name you had to guess at.

If you have not yet sat any Cloudera exam, there is a case for taking the Generalist first. It covers the whole platform at a shallower depth, and the Cloudera Generalist certification uses the same Questionmark delivery, the same 60 percent pass mark and the same two-year validity, so it doubles as a rehearsal of the exam process. It is not a prerequisite; nothing is.

How Does Exam Day Work Through Questionmark?

CDP-6001 is taken online through Questionmark with a live proctor on webcam, and there is no test-centre option. Cloudera asks candidates to install Zoom and run a system and bandwidth test 48 hours ahead, use a wired connection rather than wifi, clear the desk, stay alone and on camera for the whole 90 minutes, and open nothing but the exam window.

Results appear on screen and by email as soon as the exam is submitted, with a breakdown by topic but no view of individual questions. The Cloudera certification FAQ sets the retake terms: a failed candidate waits seven calendar days, pays the full fee for each further attempt, and a passed exam cannot be retaken to improve the score. If the connection drops, the clock keeps running while you reconnect. A no-show forfeits the fee. The pass is valid for two years, after which the exam is sat again to keep the status.

Frequently Asked Questions

Does Cloudera offer a data scientist certification?

No. Cloudera’s catalogue has no data scientist exam. The Cloudera Machine Learning Engineer certification, CDP-6001, is the credential for people who build, tune and deploy models on Cloudera, and it is the one a data scientist on the platform would take.

How much does the Cloudera Machine Learning Engineer certification cost?

CDP-6001 costs $330 USD, bought through the Cloudera Education store. A retake costs the full fee again; Cloudera offers no retake discount.

How many questions are on CDP-6001 and how long is it?

45 questions in 90 minutes, with a passing score of 60 percent, which works out to 27 correct answers.

Is CDP-6001 the same as the Cloudera CDP Machine Learning Engineer exam?

Yes. Cloudera’s catalogue now titles it Cloudera Machine Learning Engineer, the exam guide says CDP Machine Learning Engineer, and AnalyticsExam says Cloudera CDP Machine Learning Engineer. All three refer to exam code CDP-6001.

What is the biggest domain on the Cloudera ML Engineer exam?

Cloudera Machine Learning, at 31 percent, covering workspaces, projects, experiments, accelerators for ML projects, data visualizations, runtimes and GPUs. Spark MLlib is next at 22 percent.

Is Cloudera Machine Learning the same as Cloudera AI?

Yes. The exam guide uses the older name Cloudera Machine Learning; Cloudera’s documentation and product pages now call the service Cloudera AI and its workbench the Cloudera AI Workbench. The features tested are the same.

Do I need to take Cloudera training before the exam?

No. The exam guide lists Predicting with Cloudera Machine Learning and CDP Data Visualization Training as suggested courses and states they are optional. There are no certification prerequisites.

Can I take CDP-6001 at a test centre?

No. Cloudera certification exams are delivered online only, through Questionmark with a live webcam proctor.

How long is the Cloudera Machine Learning Engineer certification valid?

Two years. After it expires you may retake the exam to maintain the certification. A passed exam cannot be retaken earlier to improve the score.

Is the Cloudera certification worth it for a machine learning engineer?

It is worth it when your employer or target employer runs Cloudera, because CDP-6001 is the only machine learning credential on Cloudera’s list and its platform domain is not covered by generic ML certifications. If you do not work on Cloudera, a cloud vendor’s ML engineer credential matches more job listings.

Conclusion

The Cloudera Machine Learning Engineer certification is the answer to a search Cloudera’s catalogue cannot otherwise satisfy, and it is a narrower exam than its name suggests. CDP-6001 spends 31 percent on the workbench itself, 40 percent on Spark and MLlib, 18 percent on getting a model served and watched, and only 11 percent on the theory most candidates bring with them. Study the workbench under both of its names, write the PySpark and pipeline code yourself, deploy one model end to end, and sit the sample questions on a clock. Forty-five questions, 90 minutes, 27 right answers and two years of validity are the whole deal; the preparation that earns them is hands on, and it starts with opening a workspace rather than a textbook.

Who Needs the SnowPro Advanced Administrator Certification?

The SnowPro Advanced Administrator certification is for people who already run a Snowflake account in production and want proof that they can do it well. Snowflake asks for two or more years of hands-on administration, and the ADA-C02 blueprint spends most of its weight on access control, monitoring, cost and governance rather than on writing queries.

On its official certification page, Snowflake lists five abilities the credential validates: managing and administering accounts, data security and governance; managing and maintaining database objects and virtual warehouses; database monitoring, tuning and optimization; data sharing and use of the Snowflake Marketplace; and disaster recovery, backups and replication. It adds that successful candidates may have fluency with ANSI and Snowflake advanced SQL.

Read that list as a job description and three kinds of people fit it well:

• Platform or database administrators who own a Snowflake account and hold the ACCOUNTADMIN or SECURITYADMIN role day to day
• Data platform engineers who manage warehouses, resource monitors and cost reporting for several teams
• Security and governance leads who set up single sign-on, network policies, masking policies and tagging

It suits analysts and developers less well. Someone who mostly writes transformations would find the Data Engineer exam closer to their work, and someone who mostly builds reports would find the Data Analyst exam closer. Snowflake added the Administrator credential to the Advanced series in June 2022, and its launch announcement carried a line from a practitioner on what SnowPro credentials do for a career.

“It has helped advance my career and also helped me develop a team of engineers who are also certified SnowPros to meet and exceed our clients’ expectations to deploy Snowflake environments.”

How Is ADA-C02 Run: Questions, Time, Score and Fee?

ADA-C02 has 65 questions and a 115 minute time limit, and it costs $375 per attempt. Results are scaled from 0 to 1000, and you need 750 to pass. The exam is scheduled through Pearson VUE, and the recommended course is Snowflake’s Administering Snowflake training. The fee matches the price Snowflake shows for every SnowPro Advanced exam.

Sixty-five questions in 115 minutes leaves a little under two minutes for each one, which is enough if you know the commands and tight if you have to reason from first principles. Most items describe a business requirement and ask which grant, policy, parameter or feature meets it. Working through the ADA-C02 sample questions before you book shows that style quickly, because the wrong answers are usually real Snowflake features applied to the wrong problem.

A scaled score means the 750 is not a simple percentage of 65. Treat it as a high bar and aim to be comfortable across all six domains rather than calculating how many questions you can miss.

What Do the Six ADA-C02 Domains Weigh?

The ADA-C02 blueprint has six domains. Security, role-based access control and user administration carries 31 percent, performance monitoring and tuning 20 percent, account management and data governance 18 percent, and data and object management 15 percent. Disaster recovery takes 9 percent and data sharing 7 percent, so the first four together make up 84 percent.

The weights matter for planning, but do not skip the small domains. Data sharing and disaster recovery have short weights and long objective lists, and their questions tend to be precise: which object types replicate, what a reader account needs, what happens to a policy after failover. A handful of questions there can decide a scaled result near the line.

Why Does Security and Access Control Carry 31 Percent?

Security carries 31 percent of ADA-C02 because an administrator’s first job is deciding who can reach what. The domain covers administrative roles, access control design, privileges and grants, authentication methods, network and private connectivity, and SCIM provisioning. It is the largest domain on the exam and the one with the most objectives.

Roles, grants and the access control framework

The blueprint starts with the administrative roles. You need to know when ORGADMIN and ACCOUNTADMIN are the right tool, how to secure ACCOUNTADMIN, and how organization-level changes affect objects in each account. From there it asks you to design an access control framework from business requirements, which means knowing the difference between discretionary, role-based and user-based access control and when each applies. Snowflake’s access control overview explains how those frameworks combine in practice.

The grant objectives are detailed. Expect questions on role inheritance when privileges are granted or revoked, database roles, primary and secondary roles, future grants and their restrictions, managed access schemas, warehouse privileges such as USAGE, OPERATE, MODIFY and MONITOR, and how cloning an object affects the privileges granted on it. Auditing sits here too, through the ACCOUNT_USAGE and ORGANIZATION_USAGE schemas, along with newer items on securing access to AI features and models.

Authentication and network controls

The second half of the domain is about how people and programs sign in. It covers federated authentication and single sign-on through an identity provider, password policies, MFA enrollment and reporting on users without MFA, user types (PERSON, NULL and SERVICE), key-pair authentication and rotation, and programmatic access tokens. OAuth gets its own objective: comparing Snowflake OAuth with External OAuth and configuring Snowflake OAuth for custom clients. The underlying flows are defined in the OAuth 2.0 specification, which is worth reading once so the grant types stop looking arbitrary.

Network rules and network policies follow, including how account-level and user-level rules interact, then private connectivity to internal stages and the Snowflake service, and securing the Snowflake SQL API. The domain ends with SCIM for user and group provisioning, service accounts and API integrations, and two account parameters that prevent data exfiltration through unload and stage creation.

What Sits in Account Management, Governance and Object Management?

Account Management and Data Governance (18 percent) covers organizations, accounts, encryption and data protection policies, while Data and Object Management (15 percent) covers virtual warehouses, tables, views, stages, tasks and queries. Together they make up a third of ADA-C02 and test whether you can build and protect the objects that everything else depends on.

Organizations, encryption and governance policies

On the account side, you need to weigh the benefits and costs of a Snowflake organization, create and name accounts, find which regions an organization can use, and tell account names from account locators and region groups. Encryption appears as Tri-Secret Secure, how Snowflake encrypts customer data, and key rotation with periodic rekeying. Governance is the larger part: masking policies for column-level security, external tokenization and when to choose it over masking, row access policies compared with secure views, the ACCESS_HISTORY views, tagging with tag-based masking, data classification, Trust Center, and Horizon Catalog with Universal Search and data lineage.

Warehouses, tables and tasks

Object management starts with virtual warehouses: sizing and types, auto-suspend and auto-resume, and multi-cluster warehouses with their scaling policies. Tables include external tables and Iceberg tables, with EXTERNAL_VOLUME and the Polaris Catalog named in the staging objectives. If you have not worked with the open table format before, the Apache Iceberg project is the clearest place to learn what Snowflake is integrating with. The domain also covers storage integrations, task scheduling and permissions, user-managed versus Snowflake-managed tasks, sequences, persisted query results, cancelling statements, query history filters and Snowsight dashboards.

How Do Performance Tuning and Cost Control Show Up on the Exam?

Performance Monitoring and Tuning is 20 percent of ADA-C02, the second largest domain. It asks you to read Query Profiles, troubleshoot slow queries, compare caches, choose performance features, manage DML concurrency, set up resource monitors, use event tables for logging and tracing, and manage costs across warehouses, serverless features and AI functions.

Reading a slow query

The tuning objectives begin with the Query Profile: steps, the operator tree, operator nodes and types, and the difference between compile-time and runtime optimization. The troubleshooting items are specific. You should recognise data spilling and what fixes it, notice when pruning is not happening, and know the timeout parameters. Caching is tested three ways: the result cache, the local disk cache of a warehouse and what suspending it does, and the metadata cache. The fixes the blueprint names are materialized views, the search optimization service, the query acceleration service and external tables.

Resource monitors, logging and the bill

Cost control is where administrators earn their keep. The blueprint asks you to create resource monitors from business requirements, build a cost dashboard, use Budgets and WAREHOUSE_MONITORING, calculate and reduce idle time cost, and monitor automatic clustering. It also covers how the cloud services layer consumes credits for features such as Snowpipe, materialized views and clustering, plus AI and ML function costs and Notebooks consumption. Event tables round out the domain: default versus custom tables, log, metric and trace levels, alerts, and using trace data for root cause analysis.

Data Sharing and Disaster Recovery: Small Weights, Long Lists

Data Sharing and Snowflake Marketplace (7 percent) and Disaster Recovery, Backup, and Data Replication (9 percent) are the two smallest ADA-C02 domains, together 16 percent. Both have long objective lists, and their questions are precise about what replicates, what a share can contain and what happens to access controls when an account fails over.

The sharing domain covers sharing models (one to one, one to many, private and public), sharing across editions, regions and clouds, cross-cloud auto fulfillment for listings, secure views and secure UDFs in a share, reader accounts with their users, roles and resource monitors, inbound shares, Snowflake Data Clean Rooms, and the Marketplace from the provider side: profiles, listings, listing requests and usage monitoring, plus Native Apps.

The recovery domain is about replication groups and failover groups. You need to tell primary from secondary databases, replicate database and account-level objects, schedule replication, and judge how replication affects automatic clustering, materialized views, external tables, policies and tags, streams and tasks, stages and pipes, and Iceberg tables. Snowflake’s replication and failover introduction covers the concepts behind these items. The domain closes with client redirection after failover, business continuity planning, and Time Travel and Fail-safe: retention periods, querying historical data, restoring dropped objects and the edition limits on Time Travel.

How Does the Administrator Exam Sit Among the Other SnowPro Advanced Credentials?

SnowPro Advanced is Snowflake’s role-based series, and the Administrator credential is one of seven alongside Architect, Security Engineer, Data Engineer, MLOps Engineer, Data Scientist and Data Analyst. Each assesses advanced skills for one job role, so the right choice depends on what you do in Snowflake every day, not on which exam looks easiest.

The Administrator exam overlaps most with Security Engineer and Architect, because all three care about access control and account design. It overlaps least with the data-facing roles. A data engineer’s day is pipelines, streams and transformations, which is why the DEA-C02 data engineer path centres on data movement and transformation rather than on grants and network policies.

If you hold several roles, start with the one that matches most of your working week. For an administrator that means ADA-C02 first, because a third of the exam is the access control work you already do. Architects who also run the account sometimes take both, but the two exams test different judgement: the architect exam asks how a platform should be designed, the administrator exam asks how it is run and kept within budget.

How Should You Prepare for ADA-C02?

Prepare for ADA-C02 in the order of its weights, and do every objective yourself in a Snowflake account rather than only reading about it. Start with security and access control, then performance and cost, then governance and object management, and finish with sharing and recovery. Two years of production administration is the baseline Snowflake expects.

1. Read the current ADA-C02 objectives and mark every item you have not done yourself in a Snowflake account.
2. Build a small role hierarchy with custom roles, database roles, future grants and a managed access schema, then audit it through ACCOUNT_USAGE.
3. Set up key-pair authentication for a service user, a password policy, an MFA report and a network policy at account and user level.
4. Apply a masking policy, a row access policy and a tag-based masking policy to one table and check the result as two different roles.
5. Run a slow query, read its Query Profile, and try the search optimization service and query acceleration on it.
6. Create a resource monitor and a budget, then calculate a warehouse’s idle time cost from the usage views.
7. Create a share and a reader account, then set up a replication group and a failover group between two accounts.
8. Finish with timed practice and reread the objectives to find anything you still have not touched.

Keep your notes organised by domain, and track the newer items separately. The current blueprint names Cortex AI features, Horizon Catalog, Trust Center, programmatic access tokens and the Polaris Catalog, all of which are recent enough that older study material may not cover them. For a sense of how another advanced Snowflake blueprint distributes its marks across domains, the DAA-C01 domain weights make a useful comparison.

Frequently Asked Questions

How many questions are on the ADA-C02 exam?

ADA-C02 has 65 questions with a time limit of 115 minutes.

What is the passing score for the SnowPro Advanced Administrator certification?

You need 750 on a scaled score that runs from 0 to 1000. Because the score is scaled, 750 is not a fixed number of correct answers.

How much does the SnowPro Advanced Administrator exam cost?

The fee is $375 per attempt, the same as every SnowPro Advanced exam, and each registration requires the full amount.

How much experience do I need before taking ADA-C02?

Snowflake describes the candidate as someone with two or more years of hands-on experience as a Snowflake Administrator in a production environment, who may also be fluent in ANSI and Snowflake advanced SQL.

Do I need the SnowPro Core certification first?

The ADA-C02 page shows a candidate profile rather than a formal prerequisite list. Check the registration terms on Snowflake’s certification site before you book, since that is where any requirement is applied.

What is the difference between ADA-C01 and ADA-C02?

ADA-C01 was the earlier version of the SnowPro Advanced Administrator exam. ADA-C02 is the only version in Snowflake’s current certification catalogue, so it is the one to prepare for.

Which ADA-C02 domain carries the most weight?

Snowflake Security, Role-Based Access Control (RBAC), and User Administration carries 31 percent, followed by Performance Monitoring and Tuning at 20 percent.

Does ADA-C02 cover Iceberg tables and AI features?

Yes. The blueprint names Iceberg tables, EXTERNAL_VOLUME and the Polaris Catalog, and it includes managing access to AI features and monitoring AI and ML function costs.

Where do I schedule the exam?

ADA-C02 is scheduled through Pearson VUE.

Conclusion

The SnowPro Advanced Administrator certification belongs to whoever runs the Snowflake account: the roles, the sign-in methods, the network rules, the cost reports and the failover plan. ADA-C02 has 65 questions in 115 minutes, costs $375 and needs a scaled 750 out of 1000. Security and access control carries 31 percent and performance and cost another 20, so those two domains decide most of the result, while governance, object management, sharing and recovery fill in the rest with precise, scenario-based questions. Snowflake expects two years of production administration, and the best preparation is doing every objective yourself in an account. Start with the sample questions to see the style, then work through the blueprint in the order of its weights.

What Does the SAS Statistics for Machine Learning Certification Actually Prove?

The SAS statistics for machine learning certification proves that a candidate can reason statistically about a predictive model: describe and sample data correctly, test a hypothesis, fit and diagnose a linear regression, build and score a logistic regression, and explain the bias-variance tradeoff, regularization and cross-validation that machine learning borrows from statistics. A00-480 is the associate-level exam behind that claim.

SAS lists five skill areas on the credential page, and they match the AnalyticsExam syllabus objective for objective: statistics and machine learning, fundamental statistical concepts, explanatory modeling using linear regression, predictive modeling using logistic regression, and the statistical foundations of machine learning. Read that list again and notice what is absent. There is no data step, no CAS action, no Model Studio pipeline. The exam is deliberately platform-light, which is why the recommended course is titled Statistics You Need to Know for Machine Learning rather than anything with Viya in the name.

That makes the credential useful in a way most vendor badges are not. A hiring manager reading it learns that you can explain why a p-value matters in feature selection, why a model that scores well on training data may still be useless, and why an odds ratio of 1.0 means a predictor is doing nothing. Those are the questions that separate a data scientist from someone who can press the train button.

Who the exam is written for

The syllabus assumes a reader who already works with data and wants the statistical reasoning formalized: analysts moving from reporting into modeling, programmers who build pipelines without owning the model choices, and graduates who know the mathematics but have never had to defend a classification cutoff to a business owner. It is not an exam for statisticians who already publish, and it is not a coding exam for SAS programmers.

How Is A00-480 Delivered: Format, Pass Mark, Fee and Validity?

A00-480 is a 60 question exam with 105 minutes to complete it and a passing score of 68 percent, priced at $120 USD and delivered through Pearson VUE. SAS describes the items as multiple choice and short answer, states that the certification expires after 5 years, and bases the current paper on SAS Viya 2022.09 LTS.

Sixty questions in 105 minutes is 105 seconds each, which is generous for recall items and tight for the short-answer calculations. Expect to read an output table, a residual plot or a classification table and be asked what it shows: which predictor to drop, whether an assumption holds, what the c-statistic implies. Working through the A00-480 sample questions before you book is the quickest way to see that house style, because the item wording is closer to a statistics tutorial than to a software manual.

Two practical points come from the SAS credential page rather than from the syllabus. The fee is the same $120 in every country, and students and educators qualify for academic discounts on exams. The five year expiry also matters for planning: the Specialist credential that this exam feeds into inherits the renewal clock of its component exams, so sitting A00-480 and A00-485 close together keeps the two expiry dates aligned.

Which Five Objectives Make Up the A00-480 Syllabus?

The A00-480 syllabus has 5 objectives: Statistics and Machine Learning at 9 to 12 percent, Fundamental Statistical Concepts at 17 to 21 percent, Explanatory Modeling Using Linear Regression at 18 to 24 percent, Predictive Modeling Using Logistic Regression at 25 to 31 percent, and Statistical Foundations of Machine Learning at 18 to 24 percent. Together the two regression objectives account for roughly half the paper.

Objective 1: Statistics and Machine Learning

The smallest objective sets the vocabulary. You describe the ways of obtaining data and the different data types and how each is analyzed, define big data and the smart applications it produces, compare machine learning with classical statistics, and explain why statistics matters in machine learning at all. The terminology sub-area asks you to relate statistical words to their machine learning equivalents, compare variable types and levels of measurement, and use common modeling vocabulary. Expect questions where a target is called a dependent variable in one sentence and a label in the next.

Objective 2: Fundamental Statistical Concepts

This is the classical core. Basics of statistical analysis covers populations against samples, the process of statistical analysis, descriptive against inferential statistics, and sampling methods including event-based sampling. Descriptive statistics covers measures of central tendency, position and dispersion, how to visualize a distribution, why the normal distribution is useful in machine learning, and measures of distribution shape. Inferential statistics covers sampling distributions, confidence intervals, hypothesis tests, the one-sample t test, and why p-values are useful in machine learning.

Objective 3: Explanatory Modeling Using Linear Regression

Three sub-areas build a regression from scatterplot to diagnosis. Correlation and simple linear regression defines explanatory modeling, explores bivariate relationships with scatterplots, contrasts correlation with covariance, uses correlation to find irrelevant and redundant predictors, and explains simple linear regression with OLS estimation, hypothesis tests and model fit. Multiple regression and model selection adds categorical predictors, ANOVA and its relation to regression, interaction effects, model comparison with R-square, adjusted R-square and information criteria, and sequential selection methods. Model diagnostics covers the assumptions of linear regression, residual plots, collinearity and its remedies, and the problems caused by outliers, leverage points and influential observations.

Objective 4: Predictive Modeling Using Logistic Regression

The heaviest objective is covered in its own section below. Its four sub-areas are an introduction to predictive modeling, categorical associations, the logistic regression model, and model deployment.

Objective 5: Statistical Foundations of Machine Learning

The overview defines machine learning, distinguishes supervised, unsupervised, semi-supervised and reinforcement learning, explains neural networks, names common algorithms, and separates data preparation from data preprocessing. The pre-processing sub-area covers common modeling difficulties, the challenge of visualizing big data, diagnosing and correcting errors, missing values and outliers, why and how to transform inputs, high dimensional data and feature engineering, and feature scaling. The final sub-area, model evaluation, estimation and post-training tasks, is where signal and noise, cross-validation, bootstrap aggregation, coefficient shrinkage, L1, L2 and L12 regularizations, estimation criteria, parameters against hyperparameters, and model interpretability all live.

Why Does Logistic Regression Carry the Most Marks?

Logistic regression carries 25 to 31 percent of A00-480 because it is the bridge between the statistics in the first half of the syllabus and the machine learning in the second. It is the first model on the paper that is built to predict rather than to explain, so it introduces honest assessment, data partitioning, the bias-variance tradeoff and scoring in one place.

The objective opens with the predictive modeling ideas that every later algorithm reuses. You compare explanatory and predictive modeling, define honest assessment through data partitioning, explain how different time frames enter a predictive model, describe how to optimize model complexity for prediction, and explain the bias-variance tradeoff. None of that is specific to logistic regression, which is the point: SAS teaches the discipline on the simplest predictive model so that it carries over to trees and gradient boosting later in the track.

Categorical associations before the model

Before fitting anything you are expected to measure association between categorical predictors, define and use Cramer’s V statistic, and interpret odds ratios. The odds ratio is the single most examinable idea in the objective because it appears twice: once as a descriptive association and again as the exponentiated coefficient of the fitted model.

The model itself

The logistic regression sub-area asks you to define the model, define odds and log odds, describe maximum likelihood estimation, interpret coefficients, assess fit, use categorical predictors, explain interaction effects, compare models using concordant and discordant pairs, the c-statistic and information criteria, and describe sequential model selection. Notice the deliberate echo of the linear regression objective: categorical predictors, interactions, information criteria and sequential selection appear in both, and a well-set question will ask what changes when the target becomes binary.

Deployment is on the paper

The objective ends with model deployment: how a logistic regression model is deployed, how it is scored, and what the classification cutoff does to a scored probability. That last topic is the one candidates from a pure statistics background most often get wrong, because the cutoff is a business decision about the cost of false positives and false negatives, not a property of the model. A model with a strong c-statistic can still be deployed badly at the wrong cutoff.

Explanatory Versus Predictive Modeling: The Distinction the Exam Keeps Returning To

A00-480 draws one line through the whole syllabus: explanatory modeling asks whether a relationship is real and what its size is, predictive modeling asks whether a scored output will hold up on data the model has never seen. Linear regression is taught as the explanatory case and logistic regression as the predictive case, and the diagnostics differ accordingly.

On the explanatory side the tools are inferential. You test the regression hypothesis, read p-values for individual predictors, check residual plots against the assumptions, and look for collinearity, leverage and influence because any of them can make a coefficient estimate meaningless even when the overall fit looks fine. The NIST handbook’s treatment of model validation beyond R-square is a good independent read on exactly the point the syllabus makes: a high R-square is not evidence that the model’s assumptions hold.

On the predictive side the tools are about honesty. You partition data so that the assessment is made on observations the model did not train on, you choose complexity to minimise error on that held-out data rather than on the training set, and you compare models by concordance and the c-statistic rather than by how well they explain the sample. The same regression family therefore gets two different report cards depending on the question being asked, and A00-480 expects you to know which report card applies.

Why the exam insists on the difference

Because most real failures come from mixing the two. A predictor that is statistically significant in a sample may add nothing to out-of-sample accuracy; a model that predicts well may contain coefficients no one can interpret. The syllabus topic “compare and contrast explanatory and predictive modeling” sits at the head of the heaviest objective for that reason, and the time-frame topic beside it, which asks how different time frames enter a predictive model, is the practical form of the same idea: you train on the past and score the future.

Where Does Classical Statistics End and Machine Learning Begin on This Exam?

On A00-480 the boundary falls inside objective 5. Classical statistics supplies estimation, inference and the regression models; machine learning adds the learning process, the split between parameters and hyperparameters, resampling through cross-validation and bootstrap aggregation, and the regularization penalties that trade a little bias for a lot less variance. The exam treats these as extensions of statistics, not replacements for it.

The overview sub-area is definitional. You define machine learning, separate supervised, unsupervised, semi-supervised and reinforcement learning, explain what a neural network is, and name the common algorithms. This is the only place the syllabus goes beyond regression models, and it does so by name rather than by mechanics; you will not be asked to derive backpropagation.

Pre-processing is where the marks hide

The pre-processing sub-area is longer than it looks. It covers the common difficulties of modeling data for machine learning, why big data is hard to visualize, diagnosing and correcting errors, missing values and outliers, why input variables get transformed and the simple transformations used, the problems of high dimensional data and feature engineering as the remedy, and feature scaling. Candidates who have only ever handed clean tables to an algorithm should budget real time here, because the questions are about judgement: which transformation, which imputation, and why scaling matters to some algorithms and not to others.

Evaluation and the words that get mixed up

The final sub-area carries the vocabulary that trips up self-taught practitioners. Signal and noise. Cross-validation and bootstrap aggregation, which are resampling methods with different jobs: one estimates generalisation error, the other builds an ensemble. Coefficient shrinkage and why it helps. L1, L2 and L12 regularizations, which you should be able to tell apart by what they do to a coefficient. The learning process and estimation criteria. Parameters against hyperparameters. Model interpretability. If you want a platform-neutral reference for the resampling half, the scikit-learn cross-validation guide explains k-fold, stratified and time-series splits in the same terms the syllabus uses, and the ideas transfer to SAS Viya unchanged.

How Does A00-480 Combine With A00-485 to Make the Specialist Credential?

SAS awards SAS Certified Specialist: Statistics for Machine Learning to candidates who hold both SAS Certified Associate: Applied Statistics for Machine Learning (A00-480) and SAS Certified Associate: Modeling Using SAS Visual Statistics (A00-485). There is no separate Specialist exam. A00-480 supplies the statistical reasoning; A00-485 supplies the hands-on modeling in SAS Visual Statistics.

That stacking explains why A00-480 feels incomplete on its own. It is the theory half of a two-part credential. The Specialist credential page states the combination rule and adds one exemption: candidates who already hold SAS Certified Specialist: Visual Modeling using SAS Visual Statistics 8.4 on SAS Viya do not need to take the A00-485 exam. For everyone else the second paper is a 55 to 60 question, 110 minute exam at the same 68 percent pass mark and the same five year expiry, based on SAS Viya 2025.03 LTS, and it gives 41 percent of its marks to building and assessing regression-type models. The A00-485 exam topics are worth reading before you decide the order, because the regression you learn statistically in A00-480 is the regression you then build in Visual Statistics.

The order to sit them

Sit A00-480 first. Its syllabus is the explanation of everything A00-485 asks you to click through, and the odds ratios, c-statistics and residual diagnostics you interpret in Visual Statistics are precisely the items you are tested on here. Candidates who do it the other way round tend to pass A00-485 by pattern recognition and then struggle with the short-answer statistics on A00-480.

A00-240 is a different credential

Search traffic for the statistical business analyst credential lands on pages about this exam often enough to be worth a paragraph. A00-240 is SAS Statistical Business Analysis Using SAS 9 (Regression and Modeling), a separate credential on the SAS 9 platform. It shares the regression and modeling subject matter but it is not part of the Statistics for Machine Learning path, it is not based on SAS Viya, and passing it does not count toward the Specialist title. If your goal is the machine learning track, A00-480 and A00-485 are the two codes that matter.

Who Should Sit A00-480, and How Should You Prepare in Six Weeks?

A00-480 suits analysts and programmers who already use SAS Viya and want the statistical foundations behind its models made explicit, and it suits anyone building toward the Specialist credential or the supervised machine learning exam that follows it. Six weeks is realistic for a working candidate: two on classical statistics, two on the regression objectives, one on the machine learning foundations and one on practice.

If you already run gradient boosting pipelines and your interest is production modeling rather than statistical reasoning, read the supervised machine learning pipelines path first, because A00-406 is the credential that tests what you do all day. Most readers of this article are not in that position: they can build a model and cannot yet defend its statistics, and A00-480 is the exam that closes that gap.

A six week preparation order

1. Weeks 1 and 2: work through objective 2 in full, from populations and sampling to the one-sample t test, and write out in your own words why p-values are useful in machine learning.
2. Week 3: fit, test and diagnose a multiple linear regression with at least one categorical predictor and one interaction, then read its residual plots for collinearity, leverage and influence.
3. Week 4: build a logistic regression on partitioned data, interpret every odds ratio, compare two models by c-statistic and information criteria, and choose a classification cutoff for a stated cost of error.
4. Week 5: cover objective 5 by contrast, writing a one-line distinction for each pair the syllabus names: parameter and hyperparameter, cross-validation and bootstrap aggregation, L1 and L2 regularization, data preparation and data preprocessing.
5. Week 6: sit timed practice at 105 seconds a question, review every miss against the syllabus topic line it belongs to, and reread objective 1 vocabulary the day before the exam.

The retake rules to know before you book

The SAS certification FAQ allows up to five attempts at an exam in any 12 month period with at least 14 days between attempts, and every attempt is charged. Once you pass, no further attempts on that exam are allowed unless the credential expires, in which case the associated exams can be retaken after one year. A first attempt that misses 68 percent therefore costs you two weeks and another $120, not the credential.

Frequently Asked Questions

Is A00-480 a SAS programming exam?

No. The syllabus names no SAS procedure, interface or code. It tests statistical reasoning: sampling, inference, linear and logistic regression, honest assessment, pre-processing, regularization and cross-validation.

How many questions does A00-480 have and how long is it?

60 questions in 105 minutes, with a passing score of 68 percent. SAS describes the items as multiple choice and short answer.

How much does the A00-480 exam cost?

$120, charged in USD in every country. Students and educators qualify for academic discounts on SAS certification exams.

Does passing A00-480 make me a SAS Certified Specialist?

No. A00-480 earns SAS Certified Associate: Applied Statistics for Machine Learning. The Specialist title, Statistics for Machine Learning, is awarded once you also hold the Modeling Using SAS Visual Statistics associate credential from exam A00-485.

Which objective is worth the most on A00-480?

Predictive Modeling Using Logistic Regression, at 25 to 31 percent. The two regression objectives together are roughly half the paper.

How long is the certification valid?

SAS states that the certification expires after 5 years.

Is A00-480 the same as the SAS Statistical Business Analyst exam?

No. The statistical business analyst credential is A00-240, SAS Statistical Business Analysis Using SAS 9 (Regression and Modeling), a separate SAS 9 credential that does not count toward the Statistics for Machine Learning Specialist title.

Do I need linear algebra or calculus for A00-480?

The syllabus does not list them. It asks you to describe maximum likelihood and OLS estimation and to interpret their outputs, not to derive them.

What happens if I fail A00-480?

SAS allows up to five attempts in a 12 month period with at least 14 days between attempts, and each attempt is charged the exam fee.

Conclusion

The SAS statistics for machine learning certification is the only SAS exam that asks you to think like a statistician rather than operate a product, and that is its value. A00-480 puts half its marks on the two regression objectives, a quarter on classical statistics and a fifth on the ideas machine learning adds: honest assessment, regularization, resampling and the parameter-hyperparameter split. It costs $120, runs 60 questions in 105 minutes at a 68 percent pass mark, and stays valid for 5 years. Treat it as the first half of the Specialist credential, sit it before A00-485, and prepare in the syllabus order: statistics, linear regression, logistic regression, then the machine learning foundations. Start with the sample questions to learn the item style, then give the logistic regression objective the weeks it deserves.

What Does the Nutanix Database Automation Certification Actually Test?

The Nutanix Database Automation certification measures whether a candidate can deploy, administer, optimise and troubleshoot database workloads through Nutanix Database Service. NCP-DB covers the NDB control plane and its automation of provisioning, patching, protection, cloning and recovery across the engines NDB manages. It does not examine SQL, query tuning or engine internals; the subject is the platform that operates the databases.

Nutanix positions the credential at the professional level, one step above the Nutanix Certified Associate and alongside the other specialty NCP tracks such as Multicloud Infrastructure, Unified Storage and Cloud Native. The official page lists the typical candidate as a database administrator or developer, an automation engineer, channel partner staff who perform NDB related work, or an OEM engineer who manages databases on the Nutanix platform, whether on premises or in a public cloud.

The practical consequence is that a strong Oracle or SQL Server DBA with no NDB exposure is not ready, and a Nutanix infrastructure administrator who has never been responsible for a database is not ready either. The exam sits at the join between the two roles, which is exactly where Nutanix Database Service sits in an organisation.

What Are the NCP-DB Exam Format and Passing Score?

NCP-DB is a 75 question multiple choice exam with a 120 minute time limit, priced at $200 USD per attempt and delivered in English and Japanese. VMExam publishes the passing score as 3000 on a scale of 1000 to 6000. Nutanix names the current paper the NCP-DB 7.5 exam and recommends the Nutanix Database Management and Automation course as preparation.

Two things in that table deserve a second look. The scaled score means a raw percentage is never reported, so there is no way to translate 3000 into a number of correct answers, and any page that claims one is guessing. The absence of weightings means the four sections cannot be ranked by marks, which changes how you study: every objective group has to be covered, and the shortest route to a pass is breadth across the whole blueprint rather than depth in a favourite area.

Seventy five questions in 120 minutes gives 96 seconds per item, which is comfortable for recall but tight for the scenario questions that describe an NDB console state and ask what to do next. Working through the NCP-DB sample questions before the real sitting shows the phrasing those scenario items use, and the official Nutanix page links the exam blueprint guide as a PDF for the complete objective list.

Which Four Sections Make Up the NCP-DB Syllabus?

The NCP-DB syllabus has four sections: Deploy and Configure an NDB Solution; Monitor Alerts and Storage Usage within an NDB Implementation; Operate and Maintain an NDB Environment; and Administer an NDB Environment. Together they hold 21 objective groups, and Nutanix publishes no percentage weighting for any of them, so the exam has to be prepared as a whole.

The section names on the VMExam syllabus page and on the official Nutanix page match exactly, and the objective groups beneath them match too. What follows is the structure as both publish it.

Counting the objective groups is the nearest thing to a weighting that the blueprint offers. The operate section has seven, the administer section has six, deployment has three and monitoring has two. That does not prove the operate section carries the most questions, but it does show where the breadth of knowledge is expected, and candidates who report the paper as operations heavy are describing the blueprint accurately.

How Do Provision, Patch, Clone and Restore Fit Together in NDB?

In Nutanix Database Service the four operations form one lifecycle around a database. Provisioning creates a database server VM and instance from profiles; patching moves that server to a newer software profile version inside a maintenance window; cloning produces a copy from a Time Machine snapshot or a point in time; restore returns the source to an earlier state. The NCP-DB operate section tests all four as linked tasks.

Provisioning starts the chain. The objectives ask you to select the engine, choose a single instance or a high availability instance, pick the Nutanix cluster or clusters, select the profiles that define software, compute, network and database parameters, and name and size the database. Every one of those choices is reused later, because a clone inherits a profile and a patch targets a software profile.

Patching through software profile versions

NDB does not patch a database server directly. It publishes a new version of the software profile, either image based or as an incremental package, tests it according to the engine vendor’s process, and then applies it immediately, on a schedule, or inside a maintenance window that you created and associated with the entity. Expect questions that describe a profile version that has been created but not published, or a maintenance window that exists but is not associated, and ask why the patch did not run.

Cloning for the people who are not DBAs

Cloning is where NDB earns its keep with development teams. The blueprint asks you to determine the business requirement first, then create the clone either onto an authorised existing VM or a new target, with pre and post create scripts, from a manual snapshot or a point in time, and optionally from a remote cluster. Clones can then be refreshed on a schedule or by hand. The exam treats refresh as a first class operation, not an afterthought.

Restore as the test of everything before it

Restore is listed with three forms: from a snapshot, to a point in time, and from a remote cluster. Each depends on protection that was set up earlier through the Time Machine, which is why the next section matters more than its single objective group suggests.

What Is Time Machine and Why Does It Dominate the Operate Section?

Time Machine is the NDB component that protects a registered or provisioned database by taking snapshots and capturing transaction logs on a schedule defined by an SLA. It is what makes point in time clones and restores possible, and its dashboard is also where the monitoring section expects you to read storage utilisation. In NCP-DB, Time Machine objectives appear under both the operate and the monitor sections.

The manage Time Machine objective group asks for three skills. First, determine the requirements: how many snapshots to keep, which schedule and retention policy applies, and how much storage that policy will consume. Second, create custom SLAs when the out of the box ones do not fit. Third, manage the Data Access Management policy, which controls replication of snapshots and logs to remote clusters, and protect databases by applying an SLA and taking manual snapshots.

The mechanism behind point in time recovery is the same one every major engine uses for its own backups: a base copy plus an ordered stream of logs that can be replayed up to a chosen moment. The PostgreSQL documentation on continuous archiving describes that model for one engine; Time Machine applies it across every engine NDB manages, with the snapshot as the base copy and log catch-up as the stream. A Nutanix architecture write-up on MySQL under NDB puts it precisely.

“Each snapshot is stamped with the executed GTID set at the moment of capture. Log catchup archives every subsequent GTID in strict order.”

For the exam, the practical reading of that quote is that a point in time clone or restore needs both halves. If log catch-up has not run since the last snapshot, the only recoverable states are the snapshots themselves. Questions that describe a restore to a time between snapshots are really asking whether you understand what the log capture schedule in the SLA is for.

Storage is a Time Machine question too

The monitoring section asks you to identify source and clone size, watch storage efficiency, track the number and size of snapshots, and scale or extend storage for a database. All of those numbers surface on the Time Machine dashboard. A retention policy that keeps too many daily snapshots shows up first as storage growth, so the monitoring objectives and the SLA objectives are two views of one decision.

What Does the Administer Section Expect From Profiles, Networks and Access?

The administer section of NCP-DB covers the configuration that every provisioning and patching task depends on: five profile types, NDB software upgrades, adding Nutanix clusters, NDB-managed networks and VLANs, Active Directory integration with custom roles, and the CLI and API. It is the section that separates someone who uses NDB from someone who runs it for other people.

The profile objectives name five profile types: software, compute, network, database parameters and Windows domain. You are expected to create each, manage its status, recognise the out of the box software profiles, and replicate a software profile to a remote cluster. Because a provisioning request is assembled from profiles, a wrong or unpublished profile is the most common root cause the exam can describe without naming it.

Clusters and networks

Adding a Nutanix cluster to NDB means verifying prerequisites, registering the cluster and configuring its storage containers. Networks are either NDB-managed, where NDB hands out addresses from a pool you define, or externally managed; the objectives ask when to use an NDB-managed network, how to create one and how to add a VLAN. The multi-cluster objective in the deployment section connects here, since a clone from a remote cluster needs that cluster registered and reachable.

Access control and the API

Access control questions centre on Active Directory integration and on roles: creating custom roles, mapping users and groups, creating users and groups, and sharing entities between them. The final objective group asks you to select the right tool for a task among the console, the CLI and the REST API, and to use the API equivalent button in the console, which exposes the call behind a UI action. That button is worth practising, because it turns every manual task into an automation example, and automation is the word in the credential’s name.

Who Should Sit NCP-DB, and How Long Does It Stay Valid?

Nutanix describes the successful NCP-DB candidate as having roughly one to two years of database experience and at least six months of hands-on NDB experience, on premises or in public cloud. The credential is valid for three years for anyone certified on or after 1 August 2025, under the policy Nutanix announced on its education blog, and earning a new NCP recertifies every active NCP and NCA credential you hold.

The experience guidance is advisory rather than enforced; there is no prerequisite certification and nothing to upload before booking. In practice the six months of NDB matters more than the two years of database work, because the exam asks about console workflows, profile behaviour and SLA settings that cannot be reasoned out from general database knowledge.

The validity change is recent enough to be worth stating plainly. Nutanix announced on its certification policy update that all technical certifications moved from a two year to a three year validation period, that certifications at version 6.5 or higher were extended automatically, and that earning a new NCP recertifies all active NCP and NCA certifications. Older third-party pages still quote two years; the Nutanix post is the current position.

For someone deciding whether NCP-DB or a different Nutanix track comes first, the recertification rule is an argument for taking the associate level early. An earlier look at the Nutanix Certified Associate covers what that entry credential validates and who it suits; passing NCP-DB afterwards renews it rather than leaving two expiry dates to track.

How Should You Prepare for NCP-DB in Six Weeks?

Prepare for NCP-DB by working the blueprint in dependency order rather than section order: deploy NDB, build profiles and networks, register and provision a database, protect it with an SLA, then clone, patch and restore it. Six weeks of two to three sessions a week is realistic for someone already using NDB; a candidate new to the product needs the NDMA course first.

1. Read the official exam blueprint guide end to end and copy every objective group into a checklist, because the exam publishes no weightings and every group has to be covered.
2. Deploy an NDB VM in a lab, assign a static IP from the console, and complete the basic configuration items: language, SSL certificate, NTP and DNS from the server CLI, and SMTP.
3. Create one of each profile type, including a Windows domain profile, and replicate a software profile to a second cluster so you understand profile status and publication.
4. Register an existing database server VM and its database, then provision a new high availability instance, so you have seen both routes into NDB and their prerequisite scripts.
5. Apply a custom SLA to a database, confirm log catch-up is running, then perform a point in time clone and a point in time restore from the same Time Machine.
6. Patch the provisioned instance by creating, publishing and applying a new software profile version inside a maintenance window you created yourself.
7. Finish with alerts, storage and the API: resolve an alert, read the Time Machine dashboard for storage, and reproduce one console action through the API equivalent button and the CLI.

Each step produces a working state the next step depends on, so a failure at step four tells you something is wrong in step three. That is also how the exam’s scenario questions are constructed. If you want a second Nutanix credential in view while you plan, the recent guide to the NCP-AI certification shows how another specialty NCP divides its objectives, and the overlap in platform knowledge is larger than the different subject suggests.

Keep the official NCP-DB certification page open while you work; it carries the knowledge objectives, the exam details and the link to the exam blueprint guide, and a change to either shows up there before it reaches any third-party summary.

Frequently Asked Questions

How many questions are on the NCP-DB exam?

Seventy five multiple choice questions, with a time limit of 120 minutes. Both the VMExam syllabus page and the official Nutanix certification page publish the same figures.

What is the NCP-DB passing score?

VMExam publishes the passing score as 3000 on a scale of 1000 to 6000. The official Nutanix page does not display a passing score, and because the score is scaled there is no published way to convert it into a number of correct answers.

How much does the Nutanix Database Automation certification cost?

The exam fee is $200 USD per attempt, as stated on the Nutanix certification page. The recommended Nutanix Database Management and Automation course is priced separately.

Does NCP-DB publish domain weightings?

No. Neither Nutanix nor the VMExam syllabus page gives a percentage for any of the four sections. The operate section has the most objective groups at seven, which is the only structural signal available.

Which version of the exam is current?

The Nutanix certification page describes the current paper as the NCP-DB 7.5 exam. Check that page before booking, because it is where a version change appears first.

Do I need to know a specific database engine for NCP-DB?

No single engine is examined. The objectives are about NDB operations such as provisioning, patching, cloning, restoring and Time Machine management, which work across the engines NDB supports. Engine specific tuning is not on the blueprint.

What languages is the NCP-DB exam offered in?

English and Japanese, according to the Nutanix certification page, which also links the exam blueprint guide in both languages.

How long is the NCP-DB certification valid?

Three years for certifications earned on or after 1 August 2025, under the policy Nutanix announced on its education blog. Earning a new NCP certification recertifies all of your active NCP and NCA credentials.

Is there a prerequisite for NCP-DB?

No prerequisite certification is listed. Nutanix recommends about one to two years of database experience and at least six months with NDB, and names the NDMA course as the related training, but none of these is enforced at booking.

What is Time Machine in NDB?

Time Machine is the NDB protection component that takes snapshots and captures logs for a database according to an SLA, which is what allows clones and restores to a point in time. It is examined in both the operate and the monitor sections of NCP-DB.

Conclusion

The Nutanix Database Automation certification rewards people who can operate a database platform rather than people who can tune a database. NCP-DB asks 75 questions in 120 minutes across four sections with no published weightings, so every objective group counts, and the operate section with its seven groups is where the breadth lives. Time Machine is the concept to understand first, because cloning, restore and storage monitoring all depend on it, and profiles are the concept to understand second, because provisioning and patching are assembled from them. Work the blueprint in dependency order in a lab, keep the official Nutanix page open for the version and the blueprint PDF, and treat the three year validity and the shared recertification rule as reasons to plan NCP-DB as part of a Nutanix path rather than a standalone exam.

What does the Talend Data Quality Certified Implementer badge prove?

The Talend Data Quality Certified Implementer badge shows that its holder passed Qlik’s TDQCI exam and can use cloud-based and on-premises Talend services to monitor and assess data quality, analyze and align data for accuracy, and identify and resolve data integrity issues. It is a product credential for people who implement quality projects, not a general data management qualification.

Three names circulate for the same exam, and it helps to know all of them before searching. Qlik’s own exam page calls it Talend Data Quality Certified Implementer. The AnalyticsExam syllabus page lists it as Qlik Talend Data Quality Implementer with the code TDQCI. The description on the Credly badge page, where Qlik issues the badge, refers to passing the Talend Data Value Certified Implementer exam under a badge titled Data Quality. All three point at one exam with one topic list.

Credly rates the badge at the Advanced level and tags it with seven skills, among them Cloud Trust Score, Cloud Data Preparation, Cloud Data Stewardship, Data Quality Assessment and Data Quality Cleansing. That list is a fair summary of the exam’s shape: three cloud applications first, then the assessment and cleansing work that is done in Talend Studio.

Where it sits among the Qlik Talend exams

Qlik groups this exam with its Qlik Talend Data Integration certification exams, alongside the developer, administrator, data catalog and API services credentials. The others test how data is moved, deployed or documented. This one tests whether the data is fit to use once it arrives, which is why its syllabus spends as much time on people and rules as on components.

How long is the TDQCI exam, and what does it cost?

The TDQCI exam for the Talend Data Quality Certified Implementer credential has 45 questions, runs for 90 minutes and needs a passing score of 65 percent. Qlik’s exam details page and the AnalyticsExam syllabus page agree on all three figures. The syllabus page lists the exam price as 250 US dollars.

Forty-five questions in 90 minutes is two minutes a question, which is comfortable for recall and tight for any item that describes a dataset and asks which tool or rule fits. Sixty-five percent of 45 is a little over 29, so plan on needing 30 correct answers and treat anything above that as margin.

Two caveats come straight from Qlik. The exam content is updated periodically, and the number and difficulty of questions may change, with the passing score adjusted to keep the standard consistent. Qlik’s exam page does not display a price and links to the purchase page instead, so confirm the amount there on the day you buy. Seeing real question wording early is the cheapest way to calibrate, and the free TDQCI sample questions show how an objective such as “describe the dimensions used to calculate data trust” turns into an item.

What the exam page leaves out

Qlik’s page for this exam does not display topic weightings, so nobody outside Qlik can say how the 45 questions are split across the five areas. Any percentage you see attached to a TDQCI topic is a guess. The honest way to plan is by objective count, which the next section sets out.

Which five topic areas does the TDQCI exam cover?

The TDQCI exam covers five topic areas: working with data in Talend Cloud Data Inventory, in Talend Cloud Data Preparation and in Talend Cloud Data Stewardship, an overview of data quality, and data quality cleansing and transformations. Qlik’s exam page and the AnalyticsExam syllabus list the same 24 objectives under those five headings, word for word.

Sixteen of the 24 objectives belong to the three cloud applications, and eight to the Studio topics. That is a count of objectives, not a weighting, but it is the only published signal of where the syllabus puts its attention, and it runs against the common assumption that this is a components exam.

The objectives, area by area

The Data Inventory area asks you to define the application in context, differentiate between the supported connection types, access datasets using the data crawler and using an API, describe the advantages and disadvantages of applying semantic types across Data Preparation, Data Inventory and Data Stewardship, describe the dimensions used to calculate data trust, and use the Talend Trust Score.

The Data Preparation area covers defining the application in context, preparing data using Data Inventory and Data Preparation together, managing Data Preparation components in Talend Studio, and transforming raw data into a form suitable for analysis. The Data Stewardship area covers defining the application, differentiating between predefined roles and role-based access, managing campaigns and tasks, building resolution campaigns, and managing data quality rules.

The overview area asks you to define data quality in context, differentiate between the types of data quality analyses, monitor data movement and automate checks with business rules, and perform reconciliations between source and target. The cleansing area asks you to use the data quality components for profiling, standardization, matching and consolidation, implement data quality transformations, establish data auditing processes, and describe the use of data privacy components.

What is the Talend Trust Score, and why is it on the exam?

The Talend Trust Score is a single quality indicator in Talend Cloud Data Inventory that rates a dataset from 0 to 5 by combining several metrics. The TDQCI exam names it in two objectives, one on the dimensions used to calculate data trust and one on using the score, so it is the most directly examinable feature in the cloud half.

Qlik’s documentation groups the factors that raise or lower the score under five axes. Learn them as a set, because a question that describes a dataset and asks why its score is low can only be answered by knowing which axis each symptom belongs to.

Notice that only two of the five axes describe the data itself. Popularity, discoverability and usage describe how people treat the dataset, which is a deliberate design choice: a clean table that nobody has documented, rated or used is still hard to trust. The Talend Trust Score documentation also describes the evolution view, which plots the score over the last 7, 30 or 90 days.

“Most companies still treat data trust like an IT hygiene issue. It’s not. It’s the foundation of every AI decision a business makes.”

Semantic types tie the three applications together

A semantic type tells Talend what a column means, such as an email address or a country, instead of only what data type it holds. The validity axis depends on it, since a value can only be judged invalid against a meaning. The syllabus asks for the advantages and disadvantages of applying semantic types across all three cloud applications, so expect questions on what happens when a type is assigned wrongly as well as when it is right.

Connections, the crawler and Data APIs

Three Data Inventory objectives concern access: the supported connection types, the data crawler and API access. On the API side, Qlik’s documentation describes Data Inventory generating an API from a dataset so that applications or business partners can consume the data it contains. Read the connection, crawler and Data API sections of the Data Inventory user guide with one question in mind: when would I pick each?

How do Data Preparation and Data Stewardship share the cleanup work?

On the TDQCI exam, Talend Cloud Data Preparation is where one person cleans and reshapes a dataset through a recipe of steps, and Talend Cloud Data Stewardship is where records that need a human decision are routed to named people as tasks. The first is self-service, the second is organised collaboration, and the syllabus tests both.

Data Preparation

The four Data Preparation objectives move from definition to practice. You are expected to explain where the application fits, prepare data that was first registered in Data Inventory, and turn raw data into something an analyst can use. The fourth objective is the one Studio developers recognise: managing Data Preparation components in Talend Studio, which is how a preparation built in the browser is applied inside a Job.

Data Stewardship

Qlik’s documentation describes campaigns and tasks as the core concepts of Data Stewardship. A campaign is the unit of work for a campaign owner, who decides what the tasks are about, what stewards must do to resolve them, who takes part and what workflow they follow. A task is the unit of work for a data steward. Campaign owners and data stewards are the two predefined roles.

The syllabus singles out resolution campaigns by name with the objective “build resolution campaigns”, so know that type best. Be equally clear on the difference the exam draws between predefined roles and role-based access: the two built-in roles come ready to assign, while administrators can also create custom roles in Talend Management Console.

What does the Talend Studio half of TDQCI expect?

The Studio half of the TDQCI exam is two topic areas with four objectives each. The overview area tests whether you can pick the right kind of analysis and automate checks with business rules. The cleansing area tests the four component families Qlik names, profiling, standardization, matching and consolidation, plus auditing and data privacy.

Overview of data quality

Start with the phrase “types of data quality analyses”. Studio profiles data at different scopes, and a question will usually describe a need and ask which analysis answers it. Checking the fill rate of one column is a different job from checking that every order row points at a real customer. The remaining objectives are operational: monitoring data as it moves, turning checks into business rules that run without a person, and reconciling a source against its target after a load.

Cleansing and transformations

The four families form a natural order. Profiling shows what is wrong. Standardization puts values into one agreed form. Matching finds records that describe the same thing. Consolidation decides which values survive into the single record that is kept. An exam item that names a symptom, such as the same supplier spelled five ways, is asking which family addresses it.

The last two objectives are easy to skip and should not be. Establishing data auditing processes is about keeping evidence of what was checked and changed. Describing the use of data privacy components is about protecting sensitive values while the data stays usable. Neither needs deep build experience, and both are stated objectives.

The vocabulary in this half is not Talend’s alone. Validity, completeness and the other quality dimensions come from the wider data management discipline, and the DAMA body of knowledge is the vendor-neutral reference that treats data quality as one of its knowledge areas. Reading how the discipline frames the problem makes the “define data quality in context” objectives much easier to answer.

Who should sit TDQCI, and what experience does Qlik recommend?

Qlik recommends the TDQCI exam for people with at least six months of experience using Talend products, including experience with data preparation, data stewardship, data inventory and data analysis in Talend Cloud. It also asks for thorough knowledge of data processing and management, and familiarity with the data quality features in Talend Studio.

Those four lines describe a mixed profile. A pure Studio developer meets the last one and may miss the second. A data steward or analyst who lives in the cloud applications meets the second and may miss the last. The candidates who find the exam straightforward are usually the ones who have sat on both sides of a quality project.

The four learning paths and their length

Qlik points candidates to four learning paths, and the Credly badge page publishes a duration for each.

That is 34 hours of structured training, with 20 of them on the cloud applications and 14 on Studio. Qlik also lists a preparation training module for this exam and six documents to read: the Data Inventory user guide and its getting started guide, the Data Preparation and Data Stewardship user guides, the Talend Studio user guide, and a paper called 10 Best Practices for Successful Data Quality.

If you want the topic list and question style on one page of this site, the existing TDQCI topics and questions page is a compact companion to the official outline.

Retakes, results and the two year expiry

A Talend Data Quality Certified Implementer result is shown immediately after the TDQCI exam ends. Qlik permits three attempts per exam, each paid in full or covered by a new voucher, and every Qlik certification is valid for two years from the date of passing. Those rules are set out in Qlik’s certification policy and apply across its exams.

If you do not pass

After a first failed attempt you must wait 48 hours. After a second, the wait is 14 calendar days from the previous test date before the third and final attempt. A third failure blocks you from registering for that exam until a new version is released. Qlik enforces the waiting periods, so a retake cannot be scheduled inside them. The full wording is in the certificates and badges policy.

If you pass

Qlik no longer issues certificates. Within 24 to 48 hours of passing, Credly sends an invitation to claim the digital badge, and the result also appears in your Qlik Learning profile. Credly sends a reminder six months before the badge expires. To keep the certification you register for the current exam before the expiry date and pass it again. Qlik says exams are usually updated every three years, sooner after a major product update.

“The hardest part of AI is rarely the model. It’s trusting the data behind it. In our business, if we can’t stand behind the data, we can’t stand behind the decisions.”

Both quotations in this article come from Qlik’s Trust Score for AI announcement, which extends the score with AI-focused dimensions. The TDQCI topic list still names the original Talend Trust Score, so study the five axes above first.

How should you prepare for the TDQCI exam?

Prepare for the TDQCI exam by following the published topic list in order, cloud applications first and Talend Studio second, and by doing each objective once in the product instead of only reading about it. With 24 objectives and 45 questions, almost every objective is likely to be asked about, so there is no safe area to skip.

1. Print the 24 objectives from Qlik’s exam page and mark each one as done at work, seen once or never touched.
2. Take the Talend Cloud Trust Score learning path, then register a dataset in Data Inventory and explain its score axis by axis.
3. Work through the Data Preparation path and apply one preparation from inside a Talend Studio Job.
4. Complete the Data Stewardship path and build a resolution campaign with at least one data steward assigned.
5. Finish the Talend Data Quality Essentials path, running one analysis and one Job for each of profiling, standardization, matching and consolidation.
6. Sit timed sets at two minutes a question and trace every miss back to the objective it came from.

The first step matters more than it looks. Because Qlik publishes no weightings, your own gap list is the only sound basis for dividing study time. Most candidates find their gaps cluster in whichever half they do not use at work.

Keep the Qlik Talend Help Center open while you practise. Qlik says candidates should be comfortable searching it for documents that match the exam topics, and its Community knowledge base articles are on the recommended reading list too. The official exam details page is the place to recheck the topic list a week before you sit, since content is updated periodically.

Data quality and metadata management sit close together in real projects, and the next step for many holders is the catalog credential. The Talend Data Catalog exam page on this site outlines that syllabus.

Frequently Asked Questions

How many questions are on the TDQCI exam?

The TDQCI exam has 45 questions and a 90 minute time limit. Qlik notes that exam content is updated periodically and that the number and difficulty of questions may change.

What is the passing score for the Talend Data Quality Certified Implementer exam?

The passing score is 65 percent. On 45 questions that is a little over 29, so aim for at least 30 correct answers. Qlik adjusts the passing score when content changes to keep the standard consistent.

How much does the Talend data quality certification cost?

The AnalyticsExam syllabus page lists the TDQCI exam price as 250 US dollars. Qlik’s exam page links to the purchase page without displaying a price, and each retake needs full payment or a new voucher.

Is TDQCI a Talend Studio exam?

Only in part. Two of the five topic areas, with 8 of the 24 objectives, are Talend Studio topics. The other three areas cover Talend Cloud Data Inventory, Data Preparation and Data Stewardship.

Does Qlik publish topic weightings for TDQCI?

No. Qlik’s exam page lists the five topic areas and their objectives without percentages. Plan your study by objective count and by your own gaps.

How long is the Talend Data Quality Certified Implementer badge valid?

Two years from the date you pass. To keep it, register for the current exam before the expiry date and pass again. Credly sends a notification six months before the badge expires.

How many times can you retake the TDQCI exam?

Qlik permits three attempts per exam. You wait 48 hours after the first failed attempt and 14 calendar days after the second. After a third failure you are blocked until a new exam is released.

What experience do you need before taking TDQCI?

Qlik recommends at least six months of experience using Talend products, experience with data preparation, stewardship, inventory and analysis in Talend Cloud, and familiarity with the data quality features in Talend Studio. These are recommendations, not entry requirements.

What are the five axes of the Talend Trust Score?

Validity, completeness, popularity, discoverability and usage. Together they produce a score from 0 to 5 for a dataset in Talend Cloud Data Inventory.

Conclusion

The Talend Data Quality Certified Implementer exam is 45 questions in 90 minutes with a 65 percent passing score, and its syllabus is wider than its reputation. Three of the five topic areas test Talend Cloud Data Inventory, Data Preparation and Data Stewardship, and two test profiling, rules, cleansing and privacy work in Talend Studio. Qlik publishes no weightings, so the 24 objectives are the plan. Learn the five Trust Score axes, build one resolution campaign, run one Job for each component family, and close the gaps in whichever half you do not use at work. Then check the topic list on Qlik’s exam page once more, book the exam on Qlik Learning, and keep the two year expiry date in your calendar from the day you pass.

What does the CRM Analytics certification actually cover?

The CRM Analytics certification, exam Analytics-Con-201, tests whether you can design and implement analytics on Salesforce’s CRM Analytics and Einstein Discovery platforms from end to end. It covers enabling the product, loading and preparing data, securing datasets and apps, designing and building dashboards, and creating predictive models, across six weighted sections.

Salesforce describes the holder as someone with experience designing and implementing on both platforms in a customer-facing or internal architect role, which is the wording on the official credential page. The word consultant matters. Questions are framed as a customer requirement followed by a choice of approach, so knowing where a button sits helps less than knowing which of two valid tools fits the stated constraint.

One product, three names

Older study material describes the same product under different names. CRM Analytics was previously sold as Tableau CRM, and before that as Einstein Analytics. Search data still shows people looking for an Einstein Analytics certification and a Tableau CRM and Einstein Discovery Consultant exam. Both lead to this credential. Neither should be confused with the Tableau exams, which Salesforce also runs and which test a separate product with its own exam codes.

How many questions, how long, and what does Analytics-Con-201 cost?

Analytics-Con-201 has 65 questions, a 90 minute time limit and a 65 percent passing score, according to the VMExam syllabus page used for this article. Registration costs USD 200 and a retake costs USD 100. The exam is booked through Salesforce’s testing providers and can be taken at a test centre or online.

Which of these figures Salesforce confirms in public

Salesforce’s own pages confirm the credential name, the exam code and all six section weights, and the weights match the syllabus page to the percentage point. The question count, duration, passing score and fees sit in Salesforce’s exam guide, a help article that does not display its contents to automated readers, so those five figures here rest on the syllabus page alone. Check them against the exam guide when you book.

That caution is practical. Third-party pages disagree with one another on this exam, quoting different passing scores and time limits, and several describe a split between scored and unscored questions that no page readable for this article states. For that reason this article gives no count of correct answers needed. Treat the passing score as a percentage and aim well above it.

Where the exam is delivered

The syllabus page lists two scheduling routes, Kryterion Webassessor and Pearson VUE. Pearson VUE runs a dedicated Salesforce exam delivery page covering test centre and online appointments, and it notes that the price of each exam is shown on that certification’s own detail page. Whichever route your booking uses, Salesforce’s program agreement applies: sharing or requesting exam questions and answers can cost you the attempt and the credential.

How are the six exam sections weighted?

Analytics-Con-201 divides its marks across six sections: Data Layer at 23 percent, Analytics Dashboard Implementation at 19, Admin/Configuration at 17, Security at 16, Analytics Dashboard Design at 13 and Einstein Discovery at 12. The two dashboard sections together come to 32 percent, so 68 percent of the exam is about something other than building a dashboard.

Reading the objectives is not the same as answering them under a clock. A sensible early step is to work through a set of Analytics-Con-201 sample questions before you plan your study weeks, because the scenario style shows quickly which sections you can reason through and which you only recognise by name.

What the 32 percent figure changes

Most people come to this exam from the dashboard side. They have built lenses, arranged widgets and presented to stakeholders. The weighting says that experience is worth roughly a third of the marks. Admin/Configuration, Data Layer and Security together are 56 percent, and they are the sections a dashboard builder is least likely to have owned on a real project. Plan your hours by weight and by weakness, not by what you enjoy building.

Why does the Data Layer carry the most weight?

The Data Layer is the largest section of Analytics-Con-201 at 23 percent because every later decision depends on it. CRM Analytics does not query Salesforce records live. It works on datasets that are extracted, transformed and stored, so the exam tests how data gets in, how it is reshaped, how often it refreshes and how it is labelled.

Getting data in

Data Manager is the workspace for this section. Connections bring in Salesforce objects and data from outside Salesforce, and data sync stages that data on a schedule so that the jobs which build datasets read a prepared copy instead of extracting everything again. Salesforce’s study material for the section also covers bringing data from more than one Salesforce org and loading a CSV file, including how date values are handled on the way in.

Recipes and dataflows

Two tools turn staged data into datasets. Recipes are the visual data preparation tool, with joins, appends and transformations added as nodes. Dataflows are the older approach, defined as a set of transformation steps. The objectives name both in the same breath, and they add one item that reveals where the product is heading: delivery management strategies that include versioning and conversion. Expect scenarios that ask which tool fits, and how to refresh without colliding with platform limits.

Extended metadata

Dataset extended metadata, shortened to XMD, controls how fields appear to users: labels, value formats and colours. The objective is small on paper and frequent in practice. A dataset with raw API field names produces dashboards nobody trusts, and the exam treats fixing that at the dataset level, once, as the correct answer over patching each chart.

How is CRM Analytics security different from Salesforce sharing?

CRM Analytics security is separate from the sharing model on Salesforce records, and Analytics-Con-201 gives it 16 percent of the marks. A dataset does not inherit record access by default. Access is controlled in layers: who can use the product, who can open an app, and which rows of a dataset each person can see.

The first layer is permissions, which overlaps with Admin/Configuration: licenses and permission sets decide who can use CRM Analytics at all. The second is app sharing, where an app and the dashboards, lenses and datasets inside it are shared with users, roles or groups at a chosen access level. The third layer is the one from the opening scene, row-level security on the dataset itself.

Sharing inheritance and security predicates

The objectives name two ways to restrict rows. Sharing inheritance lets a dataset follow the sharing settings of the Salesforce object it came from. A security predicate is a filter condition written on the dataset that compares a column in each row with something about the person running the query, such as their user ID or role. Scenarios usually turn on which one fits: predicates are the tool for data that did not come from a single supported Salesforce object, or for rules the Salesforce sharing model does not express.

Two habits help here. Read each security question for the level it is asking about, because an answer that fixes app access does nothing for row access. And remember that a dashboard query runs as the viewer, so a correct chart for the builder proves nothing about what a field sales user will see.

What do the two dashboard sections ask you to build?

The two dashboard sections of Analytics-Con-201 split the work into planning and construction. Analytics Dashboard Design, at 13 percent, covers scoping requirements, layout, UX principles and templated apps. Analytics Dashboard Implementation, at 19 percent, covers queries, widget settings, interactions, compare tables, embedding, performance and release management.

Design: requirements before widgets

The design objectives start with scoping, validating and prioritising what the business has asked for, which is consultant work, not chart work. They then cover building to CRM Analytics best practices and choosing a standard templated app when one already answers the requirement. Salesforce’s study references for this section include prebuilt apps and the Event Monitoring Analytics app, so know what a template gives you and what it needs in place before it can be installed.

Implementation: queries, interactions and SAQL

Implementation is the second largest section of the exam. Its objectives cover choosing the right query type, wiring selection and result interactions so that one widget drives another, and using windowing and time series functions inside compare tables. Salesforce’s candidate profile adds that you should be able to read and write Salesforce Analytics Query Language, called SAQL, and Salesforce Object Query Language, for the cases the visual editor cannot reach.

The profile also names JSON as required knowledge. Dashboards and several other CRM Analytics assets are stored as JSON documents, and advanced changes are made by editing that definition directly. If the format is unfamiliar, the short JSON data interchange standard defines the objects, arrays and value types you will be reading, and an hour with it pays off in every implementation scenario.

Performance and release

The last three implementation objectives are operational: making dashboards actionable and available inside Lightning pages, finding slow queries with Dashboard Inspector, and managing change with versioning and Dashboard Publisher. They are easy marks for anyone who has shipped a dashboard to real users and hard ones for anyone who has only built in a practice org.

What does Einstein Discovery add to the exam?

Einstein Discovery is the predictive part of Analytics-Con-201 and carries 12 percent. It asks you to build a model from a dataset by choosing one of three prediction types, numeric, binary or multi-classification, then to interpret the results, improve the model by changing the data, place predictions on Lightning record pages, and monitor the model through its Model Card.

The section rewards judgement over mathematics. A typical scenario gives you a model with weak results and asks what to change: remove a column that leaks the outcome, drop fields that carry no signal, add data that covers a missing group, or adjust how a variable is bucketed. Salesforce’s candidate profile asks for working knowledge of statistics and the data science lifecycle, which means recognising a poor model, not deriving an algorithm.

Why the Model Card objective matters

The final objective, monitoring and interpreting a Model Card, borrows a practice from machine learning research. The idea was set out in the paper Model Cards for Model Reporting, which argued that every trained model should ship with a short record of what it is for and how well it performs.

“Model cards are short documents accompanying trained machine learning models that provide benchmarked evaluation in a variety of conditions.”

In Einstein Discovery terms, that is the place a consultant looks to judge whether a deployed prediction is still fit for use, and the exam expects you to read it and act on it. One note on vocabulary: Salesforce’s free preparation modules still describe this section with the older word story, while the current objectives say model. They refer to the same work.

Who is this credential written for?

The CRM Analytics and Einstein Discovery Consultant credential is written for people who implement analytics on Salesforce for others: consultants, analytics developers and internal architects. Salesforce’s candidate profile assumes prior experience with business intelligence, data preparation and reporting tools, plus the ability to administer, configure and secure CRM Analytics itself.

The profile, published in Salesforce’s preparation module, is unusually specific. Rewritten as a checklist, a ready candidate can:

• Design dashboards for desktop and mobile with attention to user experience
• Read and write SAQL and SOQL
• Build data preparation processes that produce and maintain datasets
• Explain master data management and data modelling at a working level
• Build and improve predictions in Einstein Discovery, including writing results back to Salesforce
• Apply basic statistics and follow the data science lifecycle
• Lead a technical project and take requirements from stakeholders

Experience administering Salesforce itself is listed as optional, which surprises people. In practice it helps a great deal, because security and Lightning page questions assume you know what a profile, a role and a permission set are.

If your work sits closer to customer data unification than to dashboards, the neighbouring Salesforce data credential may suit you better, and our breakdown of the Data 360 Consultant exam shows how differently that one is weighted.

How should you prepare for Analytics-Con-201?

Prepare for Analytics-Con-201 in the order the data moves: enable and configure the product, load and prepare data, secure it, build dashboards, then add predictions. Salesforce publishes three free preparation modules that follow the six sections, each with scenario questions and flashcards, and a practice org is essential for every stage.

1. Start with the first of Salesforce’s three modules, the official certification prep module, and note the six section weights before you study anything else.
2. Set up a practice org with CRM Analytics enabled and assign the licenses and permission sets yourself, so the Admin/Configuration objectives are things you have done.
3. Build one dataset three ways: from a synced Salesforce object, from a CSV file and from a recipe that joins two sources, then edit its extended metadata.
4. Apply row-level security to that dataset with a security predicate, log in as a second user and confirm what that user can and cannot see.
5. Build a dashboard with at least one selection interaction and one compare table, embed it in a Lightning page and check it with Dashboard Inspector.
6. Create an Einstein Discovery model on the same data, remove a weak column, compare the two versions and read the Model Card.
7. Finish with timed practice in the scenario style, and return to the section where your answers were slowest.

The sequence is deliberate. Each step produces the asset the next one needs, and security comes before dashboards because that is the order in which real projects go wrong. For a quick reference to the topic list while you work, the exam topics page on this site keeps the sections in one place under the product’s earlier name.

The syllabus page also lists two Salesforce courses as recommended training: Building Lenses, Dashboards, and Apps in CRM Analytics, coded ANC201, and Implement and Manage CRM Analytics, coded ANC301. They map neatly onto the dashboard sections and the data and security sections in turn.

Keeping the Certification Active

Passing Analytics-Con-201 is not the end of the commitment. Salesforce requires certified professionals to complete release maintenance modules for their credentials once per year, and a credential whose maintenance deadline is missed expires. The modules are completed on Trailhead and focus on what changed in the product since the last cycle.

For this credential that requirement is more than a formality. The product has been renamed twice, recipes have steadily taken over work that dataflows used to do, and the Einstein Discovery vocabulary has moved from stories to models. A consultant who passed two years ago and skipped the release notes would recognise the exam sections and misread several of the questions. Add the maintenance deadline to your calendar on the day you pass.

Frequently Asked Questions

What is the CRM Analytics certification called officially?

Its official name is Salesforce Certified CRM Analytics and Einstein Discovery Consultant, and the exam code is Analytics-Con-201. Older material calls it the Tableau CRM and Einstein Discovery Consultant or the Einstein Analytics and Discovery Consultant, after earlier names of the same product.

How many questions are on Analytics-Con-201?

The syllabus page used for this article lists 65 questions to be answered in 90 minutes. Salesforce’s exam guide is the place to confirm the current count when you book, because the public Salesforce pages readable for this article do not display it.

What is the passing score for the CRM Analytics exam?

The syllabus page gives a passing score of 65 percent. Some third-party pages quote other figures, and none of them is a visible Salesforce source, so use 65 percent as the published figure and confirm it in the exam guide.

How much does the Analytics-Con-201 exam cost?

Registration is USD 200 and a retake is USD 100, according to the syllabus page. Pearson VUE notes that the price of each Salesforce exam is shown on that certification’s detail page, and local taxes may apply.

Which section of the exam is the largest?

Data Layer is the largest at 23 percent, followed by Analytics Dashboard Implementation at 19 percent and Admin/Configuration at 17 percent. Security is 16 percent, Analytics Dashboard Design 13 percent and Einstein Discovery 12 percent.

Is the CRM Analytics certification the same as a Tableau certification?

No. CRM Analytics is built into the Salesforce platform and was once branded Tableau CRM, but the Tableau exams cover a separate product and carry their own codes. Analytics-Con-201 is the exam for CRM Analytics and Einstein Discovery.

Do you need to know SAQL to pass?

Yes, at a reading and writing level. Salesforce’s candidate profile lists competency in SAQL and SOQL, and the implementation section covers requirements the visual editor cannot meet. You also need to be comfortable reading JSON.

How much of the exam is Einstein Discovery?

Einstein Discovery is 12 percent of Analytics-Con-201. It covers choosing a numeric, binary or multi-classification prediction, interpreting and improving a model, enabling predictions on Lightning record pages and reading a Model Card.

Does the certification expire?

It stays active as long as you complete Salesforce’s release maintenance modules, which are required once per year. If the maintenance deadline is missed, the credential expires.

Conclusion

The CRM Analytics certification is an implementation exam that happens to end in a dashboard. Analytics-Con-201 puts 23 percent on the Data Layer, 17 on Admin/Configuration and 16 on Security before a single widget is placed, gives the two dashboard sections 32 percent between them, and reserves 12 percent for Einstein Discovery. The syllabus page lists 65 questions, 90 minutes, a 65 percent passing score and a USD 200 fee, and Salesforce’s own pages confirm the name, code and weights.

Study in the order the data moves, do every step in a practice org, and test your security as a second user before you trust any chart. When the sections feel familiar, time yourself on scenario questions, book the exam, and put the yearly maintenance date in your calendar the day you pass.

What is the Cloudera Generalist certification, and who is it for?

The Cloudera Generalist certification is Cloudera’s broad platform exam, code CDP-0011. It tests general knowledge of Cloudera rather than the depth of one job, and Cloudera lists five audiences for it: administrators, developers, data analysts, data engineers and data scientists. It suits an experienced professional proving platform breadth or a newcomer starting an enterprise data career.

Cloudera’s catalogue now calls the credential simply Cloudera Generalist. The money-site syllabus page and many older resources still use the earlier name, Cloudera CDP Generalist, from the period when the platform was branded Cloudera Data Platform. Both names refer to the same exam, and the code CDP-0011 has not changed.

What sets it apart is scope, not difficulty. A role exam asks how well you can do one job on Cloudera. The Generalist asks whether you understand how the pieces fit together: where a workload runs, which service owns which task, and how security and metadata follow data across environments. A data engineer who has only ever used one service will meet questions about four others. If that one job is building and deploying models, the Cloudera Machine Learning Engineer certification is the role exam to look at next.

Why one exam covers five roles

In practice, the five roles share one platform and one governance layer. An analyst querying a warehouse, an engineer scheduling Spark jobs and a data scientist training a model in Cloudera AI all inherit the same Ranger policies and the same Atlas lineage. The Generalist tests that shared ground, which is why Cloudera positions it as applicable across roles rather than as an entry rung below them.

How many questions, how long, and what does CDP-0011 cost?

CDP-0011 has 60 questions in 90 minutes, and the money-site syllabus page lists the price as 330 USD. Cloudera’s own exam guide states a pass score of 60 percent. The exam is delivered online and proctored through Questionmark, no reference material is allowed, and the certification stays valid for two years.

What the 60 percent pass mark means in questions

On 60 questions, a 60 percent pass mark works out to 36 correct answers, so you can miss 24 and still pass. That margin is generous, but it is spread across five domains of near-equal weight, so a candidate who knows two domains deeply and three not at all will not reach it. Cloudera publishes the figure on the official Generalist exam guide, alongside the topic table this article follows.

Results appear on screen and by email as soon as you submit, with a breakdown by topic. Cloudera does not let you review specific questions afterwards, so the topic breakdown is the only feedback you get if a retake is needed.

Which five domains does CDP-0011 cover, and how are they weighted?

CDP-0011 covers five domains: Cloudera Deployments at 16 percent, Cloudera Runtime Concepts at 20 percent, Cloudera on Cloud Concepts at 20 percent, Cloudera Shared Data Experience (SDX) at 22 percent, and Data Services at 22 percent. The money-site syllabus page and Cloudera’s exam guide list the same domains, sub-topics and weights.

Three things stand out in that table. No domain is above 22 percent, so there is no single area to bank on. The smallest domain, deployments, is also the one every other domain depends on, because a question about Data Hub or Replication Manager often hides a question about where the cluster runs. And Data Catalog appears twice, once under on cloud concepts and once under SDX, which is a hint that the exam treats it as both a cloud service and a governance tool.

The fastest way to see how those topics turn into questions is to work through the Cloudera Generalist sample questions, which are written as short business scenarios: a team losing capacity to other jobs, a bank replacing cron scripts, a partner needing one hardened entry point. Each one names a situation and asks which component solves it, which is the shape of the real exam.

Cloud, Base and Data Services on premises

The deployments domain asks you to tell three form factors apart. Cloudera on cloud runs in a public cloud account that Cloudera’s control plane manages. Cloudera Base on premises is the classic cluster in your own data center. Cloudera Data Services on premises brings the containerised data services to your own hardware. Cloudera’s product leadership describes the goal behind running all three in plain terms:

“Our customers no longer accept trade-offs. They want the flexibility of the cloud, the control of the data center, and the ability to scale without disruption.”

For the exam, that sentence is the domain in miniature. Expect questions that describe a constraint, such as data that must stay in a specific country or a workload that spikes at month end, and ask which form factor fits.

Why do SDX and the data services carry 44 percent of the exam?

SDX and Data Services carry 44 percent of CDP-0011 because together they describe what makes Cloudera one platform rather than a bundle of tools. SDX is the shared security, governance and metadata layer that follows data everywhere, and the five data services are the workloads that run on top of it. A generalist must know both.

The four parts of SDX

• Apache Ranger defines and enforces access policies, including row-level filtering and column masking, and keeps an audit trail of who accessed what.
• Apache Atlas records metadata, classification and lineage, so you can trace a column back to the source that produced it.
• Apache Knox is the perimeter gateway, giving outside users a single hardened entry point instead of direct access to cluster nodes.
• Cloudera Data Catalog lets users search and understand datasets across environments, built on the metadata Atlas holds.

Ranger is worth reading about at the source, because the exam tests what it does rather than how to click through it. The Apache Ranger project describes itself as a framework to enable, monitor and manage comprehensive data security across the Hadoop platform, which is exactly the role SDX gives it. Atlas sits beside it, and the Apache Atlas project site explains the classification and lineage model that Data Catalog builds on.

The classic trap is confusing the four. A question about restricting which rows a role can see is Ranger. A question about where a dataset came from is Atlas. A question about exposing services to partners over the internet is Knox. A question about finding a dataset without knowing which environment holds it is Data Catalog.

The five data services, one job each

Questions in this domain rarely ask for a definition. They describe a problem and offer four services, two of which sound plausible. The way through is to ask what the workload actually is: batch transformation, interactive SQL, row-level serving, model work or data movement.

What do the Runtime and on cloud domains ask you to tell apart?

The Runtime and on cloud domains, 40 percent of CDP-0011 together, test whether you can pick the right component for a described situation. Runtime covers what runs inside a cluster, such as YARN scheduling, replication and Virtual Private Clusters. On cloud concepts covers the services that manage environments: Management Console, Data Hub, Observability and Data Catalog.

Runtime concepts that get confused

• YARN preemption versus speculative execution. Preemption reclaims containers from a lower-priority queue when a higher-priority queue falls below its guaranteed share. Speculative execution reruns slow tasks. Only one of them solves a starved queue.
• Virtual Private Cluster versus copying data. A Virtual Private Cluster gives a team dedicated compute against the same shared data, with no copy of the data lake.
• Replication Manager versus scripts. Replication Manager replaces hand-written copy jobs between clusters with centrally managed, policy-based replication, which is the usual disaster recovery answer.
• Streams Messaging versus Stream Analytics. One moves event streams, the other processes them as they arrive.

On cloud services that get confused

Management Console is the starting point: it registers cloud credentials, creates environments and shows the health of each one. Data Hub provides clusters built from templates for a specific workload inside an environment. Cloudera Observability reports on workload performance and cost. Data Catalog is the discovery layer. Cloudera’s Management Console documentation is the clearest single place to see how those four relate, and it is one of the documentation areas Cloudera’s exam guide recommends.

A reliable test: if the scenario mentions onboarding, credentials or creating an environment, the answer is Management Console. If it mentions a cluster for a particular workload, it is Data Hub. If it mentions cost or a slow job, it is Observability.

Should your first Cloudera exam be the Generalist or a role exam?

Take the Generalist first if you are new to Cloudera, if your work touches several services, or if you are not yet sure which role you will grow into. Take a role exam first if your job is already clearly one of administration, data engineering, analysis or machine learning, because a role exam proves the depth an employer hiring for that role checks for.

Is a Cloudera certification worth it? That depends on whether Cloudera runs where you work or where you want to work. Cloudera builds its business around financial services, telecommunications, the public sector and other industries that keep data in their own data centers as well as the cloud, so the credential carries most weight with employers in those sectors. If your target employers run a different stack, the Generalist proves broad data platform literacy but carries less direct weight.

Many candidates take the Generalist and then a role exam within the same two-year validity window. If administration is the next step, the public cloud administrator exam is the natural follow-on, and the breakdown of the CDP-5001 cloud administrator exam shows how much of the Generalist’s on cloud domain carries straight over.

How does online proctoring through Questionmark work on exam day?

CDP-0011 is taken online only, through Questionmark, with a live proctor watching over webcam, microphone and desktop sharing. There is no test center option. Cloudera asks you to install Zoom and run a system test 48 hours ahead, use a wired connection rather than wifi, and keep a completely clear desk with no notes or reference material.

Before the exam

After purchase, Cloudera University emails instructions for scheduling in Questionmark. Time slots are limited each day, so book early. You can cancel or reschedule up to 24 hours before the start; a no-show forfeits the fee. The start icon becomes active 10 minutes before the appointment.

During the exam

The rules are strict and the proctor enforces them. You must be alone in a silent room, with your face clearly visible and not backlit. No headphones, phones or other devices, no food or drink, no other applications, no screenshots and no breaks. If your connection drops and you reconnect quickly you may continue, but the timer keeps running while you are offline, which is the practical reason for the wired connection.

After the exam

Results appear immediately with a topic breakdown. A failed attempt means waiting seven calendar days, counted from the day after, before you can retake, and every attempt is paid at the full price with no retake discount. A pass cannot be retaken to improve the score. The Cloudera certification FAQ sets out each of these rules, and they apply to every exam in the program, not only the Generalist.

How should you prepare for the CDP-0011 exam?

Prepare for CDP-0011 by working through the five domains in the order the platform is built: deployments first, then Runtime, then on cloud services, then SDX, then the data services. Use Cloudera’s Essentials training and documentation areas for each, then test yourself on scenario questions until you can name the right component without hesitation.

1. Learn the three form factors first, Cloudera on cloud, Cloudera Base on premises and Cloudera Data Services on premises, and be able to say which constraint points to each.
2. Work through the seven Runtime sub-topics, writing one sentence for each on the problem it solves, starting with YARN queues and preemption.
3. Map the on cloud services in order: Management Console to create an environment, Data Hub to run a workload in it, Observability to watch it, and Data Catalog to find data across it.
4. Study SDX by contrasting Ranger, Atlas, Knox and Data Catalog side by side until each scenario has exactly one answer.
5. Match each of the five data services to a workload type, using batch, SQL, serving, models and data movement as the five labels.
6. Finish with timed scenario practice at 90 seconds a question, and review every miss against the domain it came from.

Cloudera’s exam guide names Cloudera Essentials as the suggested course and lists the documentation areas worth reading: Management Console, Cloudera on Cloud, Open Data Lakehouse, Data Services, Streams Messaging, Streaming Analytics, Observability, Data Hub and governance concepts. That list maps closely to the domain table, so it doubles as a checklist.

For a compact view of the topics and question style in one place, the existing CDP-0011 topics and questions page on this site works well as a final review sheet once the documentation reading is done.

Hands-on time helps, but it is not required in the way it is for the administrator exams. Because the Generalist asks which component fits a situation rather than how to configure it, clear mental models of each service matter more than hours at a terminal.

Frequently Asked Questions

How many questions are on the Cloudera Generalist exam?

The CDP-0011 exam has 60 questions, and you have 90 minutes to answer them. Both Cloudera’s exam guide and the money-site syllabus page give the same figures.

What is the passing score for CDP-0011?

Cloudera’s exam guide states a pass score of 60 percent, which is 36 correct answers out of 60. Some third-party pages describe the exam only as pass or fail, but Cloudera publishes the 60 percent figure.

How much does the Cloudera Generalist certification cost?

The money-site syllabus page lists the exam at 330 USD. Every retake is charged at the full price, because Cloudera offers no retake discount.

Is CDP-0011 the same as the Cloudera CDP Generalist exam?

Yes. Cloudera now calls the credential Cloudera Generalist, while older resources and the money site still use Cloudera CDP Generalist. The exam code CDP-0011 is the same.

Can I take the Cloudera Generalist exam at a test center?

No. Cloudera certification exams are delivered online only, through Questionmark, with a live proctor watching by webcam and desktop sharing.

Can I use documentation during the exam?

No. Cloudera allows no resources at all during the exam, including reference material, white papers and user guides, and you may not open any other application or website.

How long is the Cloudera Generalist certification valid?

Cloudera certifications are valid for two years. When yours expires, you can retake the exam to keep your certification status.

How soon can I retake CDP-0011 if I fail?

You must wait seven calendar days, counted from the day after the failed attempt. There is no limit on attempts, but each one is paid.

Is the Cloudera Generalist certification worth it for beginners?

For someone starting an enterprise data career on Cloudera, yes. Cloudera positions the exam for both experienced professionals and newcomers, and it covers the whole platform rather than one job.

Conclusion

The Cloudera Generalist certification is the one Cloudera exam that tests the whole platform for every role. CDP-0011 is 60 questions in 90 minutes with a 60 percent pass mark, spread over five domains between 16 and 22 percent, so breadth beats depth in any single area. The marks go to candidates who can match a described problem to the right component: Ranger or Atlas, Management Console or Data Hub, Data Engineering or Data Warehouse. Learn the platform in the order it is built, rehearse scenario questions until each has one obvious answer, and set up a wired, quiet room before booking. Take it first if your work spans services or your path is still open, then add a role exam within the two-year validity window.

What does the LLM-ACA exam look like on paper?

The LLM-ACA exam has 50 questions in 60 minutes, a passing score of 80 out of 100 points and a fee of 200 USD. It is delivered in English, and Alibaba Cloud lets you choose an online or offline sitting. Seven weighted syllabus sections make up the paper, and the exam fee is non-refundable.

Sixty minutes for 50 questions works out at 72 seconds each. That is a brisk pace, and it matters because some items ask for more than one answer. A published sample question on business prompts, for example, tells you to choose three options.

The pass mark deserves equal attention. Eighty points out of 100 is a high bar for an associate credential, and Alibaba Cloud does not publish how points are split between questions. Treat every domain as worth winning, including the small ones.

Two rules on the official LLM-ACA page affect your planning. You must leave 14 days between any two professional exams, and the exam cannot be refunded once bought. Neither that page nor the syllabus page lists a prerequisite or a validity period, so check both points with Alibaba Cloud before you book.

Where do the LLM-ACA marks actually sit?

LLM-ACA divides its marks across seven sections. Prompt engineering and enhancing LLM applications lead with 20 percent each, followed by LLM concepts at 18 percent and LLM agent at 14 percent. LLM fine-tuning has 12 percent, LLM safety and compliance has 10 percent, and multimodal LLM closes the list at 6 percent.

Three groups, not seven topics

The weights fall into three natural groups. Using a model well, meaning concepts, prompts and application enhancement, holds 58 percent. Building on top of a model, meaning agents and fine-tuning, holds 26 percent. Governing and extending it, meaning safety and multimodal work, holds the last 16 percent.

Two sections come with no objectives

LLM concepts and multimodal LLM have weights but no listed objectives. Together they are worth 24 percent, so you cannot skip them. The practical answer is to test yourself early with LLM-ACA sample questions, which show the level at which concepts are asked.

Those samples are revealing. One asks which scenario is in play when a model pulls an invoice number, vendor name and due date out of emails. Another asks why a fluent answer still needs human review. Both are concept questions, and neither needs code.

Why do prompts and plugins carry 40 percent of LLM-ACA?

Prompts and plugins carry 40 percent of LLM-ACA because the credential is aimed at people who apply large language models, not people who train them. Alibaba Cloud designed it for learners from non-technical backgrounds, engineers without algorithm expertise and beginners. Writing a good prompt and extending a model with tools are the skills that audience uses first.

Prompt engineering, 20 percent

The syllabus names three objectives here: prompt techniques, prompt frameworks and using prompts in business tasks. The sample questions show what that means in practice. Few-shot prompting, for instance, is tested as a way to guide the model by showing example inputs and outputs.

A strong business prompt, according to the answer key of the same sample set, contains three things:

• A clear task objective
• Relevant background context
• The expected response format

The wrong option in that question is worth remembering too. Unrestricted creative assumptions do not improve a business prompt. The exam rewards structure and limits, not freedom.

Enhancing LLM applications, 20 percent

This section moves from the prompt to the application around it. Its three objectives cover automating tasks with LLM APIs, how LLM plugins work, and the plugins supported on Alibaba Cloud Model Studio. That last objective is platform knowledge, so general reading about tools will not be enough.

Spend time in the Model Studio documentation before the exam. Alibaba Cloud lists knowing which of its services help you apply large models as one of the three takeaways of the certification, so expect service recognition questions. One published sample asks which storage service suits the source documents of a knowledge application, and the answer is OSS.

What does the LLM agent domain expect you to know?

The LLM agent domain is worth 14 percent of LLM-ACA and has four objectives: the principles and structure of an LLM agent, the principles and structure of ModelScope-Agent, multi-agent concepts and examples, and iteratively improving agent systems. It tests whether you understand how an agent decides and acts, not whether you can program one.

How an agent works

Two sample questions show the depth. The first asks what an observation is for in an agent loop. It provides the result of a tool action so the agent can make its next decision. The second asks why tool descriptions must be clear. They help the model choose the correct tool and supply valid inputs.

That gives you a simple model to revise from:

1. The agent reads the request and the descriptions of the tools it can use.
2. It chooses a tool and takes an action with valid inputs.
3. It reads the observation, which is the result of that action.
4. It decides what to do next, or returns an answer.

ModelScope-Agent by name

The syllabus names one framework directly. ModelScope-Agent is an open source agent framework from the ModelScope community, and its project repository on GitHub is now titled MS-Agent. Reading the project overview is a quick way to see the structure the objective refers to.

Multi-agent systems and improvement

The remaining objectives cover systems where several agents share a task, and the habit of improving an agent system in rounds. For an associate exam, focus on why work is divided between agents and how you would tell that a change made the system better.

How much fine-tuning, safety and RAG is on LLM-ACA?

On LLM-ACA, fine-tuning is worth 12 percent, LLM safety and compliance is worth 10 percent and multimodal LLM is worth 6 percent. Retrieval-augmented generation has no section of its own on the Associate syllabus, although Alibaba Cloud names a RAG question and answer bot in the certification takeaways and RAG appears in the published sample questions.

Fine-tuning is about judgement

The three fine-tuning objectives are concepts and role, key factors and when to use fine-tuning, and process and principles. Notice the middle one. The exam wants to know whether you can tell when fine-tuning is the right choice. A sample question reinforces this by asking why a fine-tuned model is compared with a baseline: to verify that fine-tuning actually improved the target task.

Safety and compliance

This section lists LLM risks and cases, risk governance recommendations, and application filing for LLM apps. The samples follow the same line. Before an external launch, a team should review safety, privacy, compliance and misuse risk. Human review matters because output can be fluent and still be inaccurate, incomplete or unsafe.

Where RAG fits

Retrieval-augmented generation lets a model answer from documents you supply. On this exam it shows up as a concept. One sample describes an answer that cites a document section which does not support the claim, and asks you to name the problem: poor faithfulness to the retrieved evidence. Learn the idea and its failure modes. The engineering detail belongs to the Professional tier.

LLM-ACA, LLM-ACP or ACA Generative AI Engineer?

LLM-ACA, LLM-ACP and ACA Generative AI Engineer all cost 200 USD, but they differ in length, pass mark and depth. LLM-ACA has 50 questions in 60 minutes at 80 points. LLM-ACP has 75 questions in 120 minutes at 80 points. ACA Generative AI Engineer has 50 questions in 90 minutes at 70 points.

Associate against Professional

The audience is the clearest difference. Alibaba Cloud describes the LLM-ACP certification as tailored for generative AI enthusiasts and application developers with a programming foundation. The Associate tier, in contrast, is open to people with no technical background.

The syllabuses confirm it. LLM-ACP gives retrieval-augmented generation its own 20 percent section and names LlamaIndex and the RAGAS framework. It also raises fine-tuning to 16 percent and adds a production section covering deployment and vLLM. None of those names appear on the LLM-ACA syllabus.

LLM Engineer against Generative AI Engineer

ACA Generative AI Engineer is the broader and gentler exam. Its syllabus has three sections: Basics of AI at 32 percent, AI Technologies at 44 percent and Implementing AI at 24 percent. You get 108 seconds per question and need 70 points. This site’s ACA-GenAI exam topics page covers that credential in detail.

LLM-ACA is narrower and stricter. It stays on large language models, gives you 36 seconds less per question and asks for 10 more points. Choose it when your work is already about prompts, plugins and agents.

How should you prepare for LLM-ACA?

Prepare for LLM-ACA in the order of its weights. Start with LLM concepts, then prompt engineering and application enhancement, which together hold 58 percent of the marks. Move on to agents, then fine-tuning, safety and multimodal models. Finish with timed practice at 72 seconds per question, because the pass mark of 80 points leaves little room.

1. Read the seven domain weights and rate your own confidence against each one before opening any material.
2. Learn the core LLM concepts first, since 18 percent of the marks depend on them and every later domain assumes them.
3. Practise writing prompts with a stated objective, background context and response format until the structure is automatic.
4. Work through how LLM APIs and plugins extend a model, and note which plugins Alibaba Cloud Model Studio supports.
5. Study the agent loop of action and observation, then read how multi-agent systems divide work.
6. Cover fine-tuning, safety and compliance, and multimodal models last, in that order of weight.
7. Sit timed sets of 50 questions in 60 minutes and review every miss by domain.

Alibaba Cloud sells an official LLM Engineer (Associate) Exam Preparation Course. It is an English e-learning course of 4 chapters covering 15 lessons, priced at 199 USD. The course is optional, since no source lists it as a requirement for the exam.

Whatever material you use, keep a simple log of missed questions by syllabus section. With seven sections and a high pass mark, a weak 6 percent domain can still decide the result.

Frequently Asked Questions

How many questions are on the LLM-ACA exam?

The LLM-ACA exam has 50 questions and lasts 60 minutes. That gives you 72 seconds per question on average. Some questions ask you to choose more than one answer, so read the instruction in each item before you respond.

What is the passing score for LLM-ACA?

The passing score is 80 out of 100 points. Alibaba Cloud does not publish how the points are divided between questions, so you cannot convert the pass mark into an exact number of correct answers. Aim to be strong in all seven sections.

How much does the LLM-ACA exam cost?

The exam fee is 200 USD. Alibaba Cloud states that the exam is non-refundable, so confirm your date and your test type before you pay. The optional official preparation course is sold separately at 199 USD.

Which LLM-ACA domain has the highest weight?

Two domains share the top spot. Prompt engineering and enhancing LLM applications are each worth 20 percent. LLM concepts follows at 18 percent, so these three sections together account for 58 percent of the marks.

Is coding required for LLM-ACA?

Alibaba Cloud says the certification is designed for people from non-technical backgrounds, engineers without algorithm expertise and beginners. The published sample questions test concepts and judgement. A programming foundation is named for the Professional tier, LLM-ACP, not for the Associate tier.

Is RAG covered on the LLM-ACA exam?

RAG has no section of its own on the LLM-ACA syllabus. However, Alibaba Cloud names a RAG question and answer bot among the certification takeaways, and a published sample question tests answer faithfulness in a RAG system. Learn the concept and its common failures.

What is the difference between LLM-ACA and LLM-ACP?

LLM-ACA has 50 questions in 60 minutes and targets beginners. LLM-ACP has 75 questions in 120 minutes and targets developers with a programming foundation. Both need 80 points and cost 200 USD. LLM-ACP adds a 20 percent RAG section and a production practices section.

Can I take the LLM-ACA exam online?

Yes. Alibaba Cloud states that you enter its exam platform and choose either the offline or the online test. The exam is available in English. You must also leave 14 days between any two professional exams.

Does LLM-ACA have prerequisites or an expiry date?

Neither the official certification page nor the syllabus page lists a prerequisite or a validity period for LLM-ACA. Because no source publishes those details, confirm them directly with Alibaba Cloud before you rely on the credential for a job application.

Conclusion

LLM-ACA is 50 questions in 60 minutes with a pass mark of 80 points, and its syllabus is unusually honest about what it values. Fine-tuning has 12 percent. Prompts and application enhancement have 40 percent between them. The exam measures how well you use a large language model.

So study in that proportion. Get the concepts firm, practise structured prompts, learn how plugins and agents extend a model on Alibaba Cloud, and then cover fine-tuning, safety and multimodal work. Time yourself at 72 seconds per question from the first practice set.

If you are still choosing between tiers, the Alibaba certification hub on this site lists the other Alibaba Cloud exams. Pick the one that matches the work you already do, and let the published weights set your plan.

What does the A00-282 exam look like on the day?

A00-282 is 60 to 70 questions in 110 minutes, and you need 68 percent to pass. The fee is 180 USD, charged in US dollars in every country, and the exam is administered by SAS and Pearson VUE. Questions mix multiple choice with short-answer items across nine weighted domains.

The short-answer items deserve a moment of thought before exam day. A multiple choice question lets you eliminate wrong options; a short-answer item does not, so a half-remembered PROC option or a hazy macro trace is harder to rescue. Candidates who practise by reading code and predicting its output, rather than by recognising answers, are better placed for both formats.

A 68 percent bar across nine domains also leaves limited room for a domain you decided to skip, because every one of them is weighted and none is marked optional.

SAS keeps every current figure on its official credential page, including the registration link that uses the A00-282 exam ID.

Which SAS credential do you need before A00-282?

You cannot sit A00-282 cold. SAS requires you to hold either the Base Programming Specialist credential or the Advanced Programming Professional credential first. It also asks for the recommended training or equivalent experience, and at least six months working as a clinical trials programmer using SAS software.

That prerequisite shapes what the exam leaves out. Because every candidate has already proved core DATA step and procedure skills, A00-282 does not spend questions on reading raw files or basic syntax. It assumes them, and builds clinical problems on top.

The six months of experience matters for a different reason. Many objectives describe tasks you only meet on a real study: deriving programming requirements from a statistical analysis plan and an annotated case report form, programming for both safety and efficacy data, or working out why two validation programs disagree. Reading about those tasks is possible. Recognising them in a scenario question is much easier if you have done them.

• Hold one of the two prerequisite SAS credentials before registering
• Have at least six months as a clinical trials programmer using SAS
• Complete the recommended SAS training, or bring equivalent experience

How are the nine A00-282 domains weighted?

A00-282 spreads nine domains between 5 and 20 percent. Validate Clinical Trial Data Reporting leads at 20 percent. Three programming domains, transform or summarise, statistical procedures and macros, carry 15 percent each. Data structures and reporting take 10 percent each, and three smaller domains take 5 percent apiece.

Read the table in two groups. The three 5 percent domains total 15 percent, less than validation alone, and they are mostly knowledge you can learn in a few focused sessions. The three 15 percent domains total 45 percent and are skills, which take practice at a keyboard rather than reading.

That makes the split of study time fairly clear. Give validation and the programming middle the bulk of it, and treat the small domains as quick wins rather than optional extras. Once you know the shape, a set of A00-282 sample questions shows how each domain is actually phrased.

Why is validation the largest domain?

Validation is 20 percent of A00-282 because clinical output is checked before anyone trusts it. The domain tests validation principles in the clinical trial industry, using the SAS log, PROC COMPARE and the MSGLEVEL option, explaining why two independent programs disagree, and identifying and resolving data, syntax and logic errors.

Two programs, one answer

The industry habit behind this domain is independent programming. One programmer produces a data set or table, a second writes their own program from the same specification, and the two results are compared. The exam asks you to determine why two independent validation programs led to a different result, which is a diagnostic question: a different derivation rule, a sort order, a missed record, or a specification read two ways.

What PROC COMPARE does not see

PROC COMPARE is the workhorse, and the blueprint is precise about its limits. You are expected to identify elements that are not validated by a comparison, such as titles, footnotes, and attributes like formats or labels, depending on how they were added to a report. A clean comparison of the data does not prove the finished table is right.

The log as evidence

The log is a validation tool in its own right. Notes about uninitialised variables, automatic type conversion or merges with repeated BY values often point to logic errors that produce plausible output. Setting MSGLEVEL to show more detail makes those notes easier to find, and the exam expects you to read them as evidence rather than noise.

• Data errors: values outside expected ranges, missing where zero was meant
• Syntax errors: code that stops or warns in the log
• Logic errors: code that runs cleanly and returns the wrong answer

The third category is the one to practise. It never announces itself, which is exactly why a second program exists.

How much CDISC does the exam expect?

A00-282 expects working knowledge, not implementation depth. The data structures domain, 10 percent, asks you to identify clinical domains, key CDISC principles and terms, the structure and purpose of SDTM and ADaM, and to trace data from raw collection through to any mapped domain. Regulatory submissions adds define.xml and SAS V5 requirements.

SDTM organises collected trial data into standard domains, and CDISC notes that it is one of the required standards for data submission to the FDA in the United States and PMDA in Japan. The CDISC SDTM standard is the reference for domain structure. ADaM sits downstream, holding the analysis-ready data that tables and figures are built from.

“ADaM standards are developed to align with analysis needs, enable analysis-ready datasets, and support the needs of reviewers, including traceability back to SDTM.”

Traceability is the thread connecting this domain to validation. A reviewer should be able to follow any result back through ADaM and SDTM to the raw data, and the exam’s tracing objective tests that you can do the same.

The submission details

Regulatory submissions is only 5 percent, but both objectives are specific. SAS V5 transport requirements constrain exported data sets, so you should know what they mean for variable names, labels and lengths. The Define-XML specification describes the metadata document that tells a reviewer what each data set and variable contains, and the exam asks for its contents and purpose.

The programming middle: transform, statistics and macros

Three domains at 15 percent each make up 45 percent of A00-282. They cover transforming or summarising clinical data, applying statistical procedures to it, and automating repeated work with macros. These are hands-on skills, and the questions tend to present code and ask what it produces or what is wrong with it.

Transform or summarise

This domain is where clinical conventions meet the DATA step. Expect categorisation and windowing of visits, storing dates to the ISO 8601 standard and checking that existing dates comply, reshaping data with PROC TRANSPOSE or with arrays, counting events, and FIRST. and LAST. processing. Calculating change from baseline is named explicitly, and it is a derivation worth practising until the edge cases, such as a missing baseline, are second nature.

Statistical procedures

You are not tested as a statistician. You are tested as the programmer who runs FREQ, UNIVARIATE, MEANS and SUMMARY correctly, knows whether a procedure suits categorical or quantitative data, spots syntax or semantic errors in FREQ, TTEST, GLM and REG code, and captures results with ODS OUTPUT or OUTPUT statements. The blueprint also asks you to program for both safety and efficacy data.

Macros

The macro domain covers creating macro variables and parameters, using automatic and user-defined variables, and wrapping repeated tasks in macros. Debugging carries weight: MPRINT, SYMBOLGEN and MLOGIC each show a different view of what the macro processor did, and questions often hinge on which one answers the problem in front of you. The SAS documentation hub covers each option in detail.

How much PROC REPORT and ODS does A00-282 test?

Reporting is 10 percent of A00-282, but its objective list is one of the longest on the blueprint. It covers PROC REPORT for tables and listings, ODS statements for destinations and output objects, and SGPLOT or SGPANEL graphs. The questions reward knowing specific statements and options rather than general familiarity.

For PROC REPORT, the blueprint names the COLUMN statement for selecting, aliasing and nesting items, DEFINE for usage, style and sort order, BREAK and RBREAK for summary rows, COMPUTE blocks, LINE statements, CALL DEFINE and the OUT= option. That is most of the procedure, so a candidate who only ever edits an inherited template will find gaps.

The ODS side asks for ODS TRACE to find output object names, SELECT and EXCLUDE to control them, and opening and closing destinations such as PDF, RTF, POWERPOINT, LISTING and WORD with their settings. Graphics include bar charts, box plots, histograms, scatter and series plots, high-low and waterfall charts, with grouping, overlays, reference lines, axes and legends.

This site has covered the credential before, and an earlier look at the clinical trials programming path covers the career side of this work in more depth than the exam facts here.

How should you prepare for A00-282?

Prepare in proportion to the weights, at a keyboard, on clinical-shaped data. Confirm your prerequisite credential first, then give validation and the three 15 percent programming domains most of your time, and close with the smaller knowledge domains and a timed run through sample and practice questions.

1. Confirm you hold the Base Programming Specialist or Advanced Programming Professional credential before you book.
2. Read the exam content guide once end to end, so every objective is familiar before you study any of them.
3. Practise validation by writing a second, independent program for a table and resolving every difference PROC COMPARE reports.
4. Work through transform, statistics and macro tasks on data with visits, baselines and adverse events, since those three domains are 45 percent of the paper.
5. Build one full listing with PROC REPORT and ODS, including COMPUTE blocks, a summary row and output to RTF or PDF.
6. Spend focused sessions on CDISC, define.xml, SAS V5 transport rules and DICTIONARY tables.
7. Finish with a timed practice exam and review every miss by domain.

SAS gathers its free practice exam, sample questions, content guide and recommended courses on the A00-282 certification pathway. The recommended reading there includes titles such as SAS Programming in the Pharmaceutical Industry, Implementing CDISC Using SAS, and Validating Clinical Trial Data Reporting with SAS, the last of which maps directly onto the largest domain.

If the prerequisite is still ahead of you, start there. The site’s guide to the A00-231 base programming exam covers the credential most candidates use to qualify.

Frequently Asked Questions

How many questions are on the A00-282 exam?

Between 60 and 70 questions, a mix of multiple choice and short-answer items, answered in 110 minutes.

What is the passing score for A00-282?

You need 68 percent to pass, as published by SAS and by the exam syllabus.

How much does the SAS clinical trials certification exam cost?

The fee is 180 USD, charged in US dollars in every country.

Do I need another SAS certification first?

Yes. You must hold either the Base Programming Specialist or the Advanced Programming Professional credential before taking A00-282.

How much experience does SAS recommend?

At least six months as a clinical trials programmer using SAS software, along with the recommended training or equivalent experience.

Which domain carries the most weight?

Validate Clinical Trial Data Reporting at 20 percent, the largest of the nine domains.

Does A00-282 test CDISC standards?

Yes. The data structures domain covers CDISC principles and terms and the structure and purpose of SDTM and ADaM, and the regulatory domain covers define.xml.

How do I register for the A00-282 exam?

Register with the exam ID A00-282. The exam is administered by SAS and Pearson VUE, and registration starts from the SAS credential page.

Is A00-282 based on SAS 9.4 or SAS Viya?

SAS 9.4. The exam name is Clinical Trials Programming Using SAS 9.4, and the credential carries the same version.

Conclusion

A00-282 is a clinical programmer’s exam with a checker’s centre of gravity. Validation is 20 percent, the largest of nine domains, and the three programming domains add 45 percent more. The format is 60 to 70 questions in 110 minutes at 68 percent, for 180 USD, and you need a prerequisite SAS credential before you can book.

Spend your time where the weight is. Write second programs and chase every PROC COMPARE difference, practise change from baseline and ISO 8601 dates until they are routine, and build a full PROC REPORT listing through ODS. Then close the smaller CDISC and submission domains, which are quick to learn and costly to leave empty.

When the domains feel familiar, test yourself under time and review each miss by domain, so the last week goes to the areas that actually cost you marks.