Databricks Certifications: Exams, Costs and Career Paths
Data Engineer, Data Analyst, Machine Learning, Generative AI, Context Engineer, or Apache Spark? Here is what each of the eight current Databricks certifications covers, how long each exam is, what it costs, and how to pick the one that fits the work you do.
Pick the certification that proves work you actually do, or work you are about to start. A badge that does not match your job helps no one. Everything below comes from Databricks’ own exam pages and FAQ, checked on 7 November 2026.
1. Choose by the work you do
Start with the tasks you want to prove you can do. Databricks groups its certifications by role: Data Analyst, Data Engineer, Machine Learning, Generative AI Engineer, Context Engineer, and Spark Developer. Only the Data Engineer and Machine Learning tracks have both an Associate and a Professional exam.
Data Analyst Associate
Data Engineer Associate
Data Engineer Professional
Machine Learning Associate
Machine Learning Professional
GenAI Engineer or Context Engineer
| If your week looks like this | Start with |
|---|---|
| Loading data, building and scheduling pipelines, fixing failed jobs | Data Engineer Associate |
| Writing SQL in Databricks, building dashboards, maintaining Genie spaces | Data Analyst Associate |
| Training models, tracking experiments, deploying endpoints | Machine Learning Associate |
| Building RAG chains and LLM apps and deploying them with Model Serving | Generative AI Engineer Associate |
| Designing what an agent sees at run time: prompts, retrieval, memory, tools | Context Engineer Associate |
| Writing Spark DataFrame code in Python and tuning it | Associate Developer for Apache Spark |
| Running production pipelines or ML systems and owning their reliability | The matching Professional exam |
2. The eight certifications
Advertisements
Databricks lists eight proctored certifications. Each card summarises the exam page. Open the linked page before you book, because Databricks updates them.
Databricks Certified Data Analyst Associate
Introductory data analysis with Databricks SQL: managing and discovering data in Unity Catalog, importing data, running and optimising queries, analysing query behaviour, dashboards and visualisations, AI/BI Genie spaces, data modelling, and securing data. The SQL on the exam follows ANSI SQL, and you do not need Python.
Best for: BI analysts, SQL analysts, and analytics developers.
Databricks Certified Data Engineer Associate
Foundational data engineering on the platform: ingestion and loading, transformation and modelling, Lakeflow Jobs, CI/CD, troubleshooting and optimisation, and governance. Code appears in SQL where possible, and in Python otherwise.
Best for: engineers new to Databricks and engineers who want their foundations checked.
Databricks Certified Data Engineer Professional
Advanced, production-level data engineering. Its scope includes data sharing and federation, which the Associate exam no longer covers, along with processing, quality, monitoring, cost and performance, security, and deployment.
Best for: engineers who design, run, and tune production workloads.
Databricks Certified Machine Learning Associate
Basic machine learning on Databricks: platform ML tooling, AutoML, MLflow, data exploration and feature engineering, training, tuning, evaluation, and deployment.
Best for: practitioners starting to build ML workflows on Databricks.
Databricks Certified Machine Learning Professional
ML engineering at production scale: scalable pipelines, distributed training, advanced MLflow, feature pipelines, MLOps, automated retraining, drift monitoring, deployment, and rollout.
Best for: experienced ML engineers who own systems after they ship.
Databricks Certified Generative AI Engineer Associate
Designing and building LLM solutions: breaking a requirement into tasks, picking models and tools, AI Search for semantic similarity, Model Serving, MLflow, and Unity Catalog governance. The ML code on the exam is Python.
Best for: developers building RAG applications and LLM chains.
Databricks Certified Context Engineer Associate
Designing and governing what AI agents receive at inference time: system prompts, retrieval with AI Search, memory with Lakebase and MLflow, tools through MCP, context window trimming and compaction, multi-agent workflows, evaluation, and governance through Unity Catalog. The code is Python. The catalog text calls it Context Engineering Associate.
Best for: engineers who build the information environment around an agent, and not only its prompts.
Databricks Certified Associate Developer for Apache Spark
Spark architecture and the DataFrame API: selecting, filtering, aggregating, joining, reading and writing, UDFs, Spark SQL, shuffles, Structured Streaming, Spark Connect, the Pandas API on Spark, and basic tuning. All code is Python.
Best for: developers who want Spark fundamentals that carry across environments.
3. Comparison table
| Certification | Level | Scored questions | Time | Fee | Languages | Code on the exam |
|---|---|---|---|---|---|---|
| Data Analyst | Associate | 45 | 90 min | USD 200 | English | ANSI SQL |
| Data Engineer | Associate | 45 | 90 min | USD 200 | 4 | SQL where possible, otherwise Python |
| Data Engineer | Professional | 59 | 120 min | USD 200 | 4 | Check the exam guide |
| Machine Learning | Associate | 48 | 90 min | USD 200 | 4 | Check the exam guide |
| Machine Learning | Professional | 59 | 120 min | USD 200 | English | Check the exam guide |
| Generative AI Engineer | Associate | 45 | 90 min | USD 200 | 4 | Python for ML, SQL for some data work |
| Context Engineer | Associate | 45 | 90 min | USD 200 | English | Python, SQL for some data work |
| Associate Developer for Apache Spark | Associate | 45 | 90 min | USD 200 | English | Python |
“4 languages” means English, Japanese, Brazilian Portuguese, and Korean. All eight exams are multiple choice, allow no test aids, run online or at a test center, have no formal prerequisite, and stay valid for two years. Databricks recommends about six months of hands-on experience with the tasks in the exam guide. Exams can include unscored questions that are not marked. For the GenAI and Context Engineer exams, Databricks says the number of questions you actually see can be higher than the stated total when unscored items are present, and extra time is already built in.
4. What each exam weighs
Domain weights tell you where to spend your study hours. They come from the exam pages. They do not tell you the exact question count per domain on your own form.
Data Analyst Associate
| Domain | Weight |
|---|---|
| Executing queries with Databricks SQL and SQL warehouses | 20% |
| Creating dashboards and visualisations | 16% |
| Analysing queries | 15% |
| Developing, sharing, and maintaining AI/BI Genie spaces | 12% |
| Understanding the Databricks Data + AI Platform | 11% |
| Managing data | 8% |
| Securing data | 8% |
| Importing data | 5% |
| Data modelling with Databricks SQL | 5% |
Generative AI Engineer Associate
| Domain | Weight |
|---|---|
| Application development | 30% |
| Assembling and deploying apps | 22% |
| Design applications | 14% |
| Data preparation | 14% |
| Evaluation and monitoring | 12% |
| Governance | 8% |
Associate Developer for Apache Spark
| Domain | Weight |
|---|---|
| Developing DataFrame and Dataset API applications | 30% |
| Apache Spark architecture and components | 20% |
| Using Spark SQL | 20% |
| Troubleshooting and tuning DataFrame applications | 10% |
| Structured Streaming | 10% |
| Using Spark Connect to deploy applications | 5% |
| Using the Pandas API on Apache Spark | 5% |
Data Engineer exams
Data Engineer Associate has seven domains, led by Data Transformation and Modeling (22%) and Data Ingestion and Loading (21%). We cover them in detail in the Data Engineer Associate guide. One detail matters when you choose between the two levels: Data Sharing and Federation is a domain on the Professional exam, at 5%, and is not in the current Associate outline.
5. Certification paths
These are suggested sequences, not requirements. Databricks lists no formal prerequisite for any of these exams.
Data engineer
Workspace, lakehouse ideas, compute, and governance vocabulary.
Ingestion, transformation, Lakeflow Jobs, CI/CD, and security.
Quality checks, orchestration, monitoring, and access control.
When you can design and debug production workloads.
Analyst or analytics engineer
Catalogs, schemas, tables, permissions, SQL warehouses.
SQL, query analysis, dashboards, Genie, modelling.
Documented SQL, access design, and business definitions.
Machine learning engineer
Features, training, evaluation, and experiment tracking.
Basic Databricks ML workflows and deployment.
Automated training, testing, monitoring, and rollout.
When you operate ML systems at scale.
GenAI application engineer
Embeddings, retrieval, prompting, evaluation, and safety.
Build, assemble, deploy, govern, and evaluate LLM apps.
Document ingestion, retrieval checks, serving, and tracking.
Context and agent engineer
Agent parts and tool use.
Retrieval, memory, MCP, context windows, evaluation, governance.
Measure retrieval quality, memory behaviour, and tool permissions.
Spark developer
DataFrames, Spark SQL, shuffles, partitioning, streaming.
Architecture and the DataFrame API.
Use the same concepts in your team’s environment.
6. Associate or Professional
Advertisements
Data Engineer and Machine Learning are the only tracks with a Professional exam. The other six certifications are Associate-level.
Foundational, task-level platform work
Production design, tuning, and operations
Databricks lists no formal prerequisite. The Associate exam is a sensible step, not a gate.
| Associate | Professional | |
|---|---|---|
| Scope | Core tasks and correct use of platform features. | Production design, reliability, optimisation, debugging, and operations. |
| Data Engineering | Ingestion, transformation, Lakeflow Jobs, CI/CD basics, monitoring, governance. | Adds data sharing and federation, plus deeper processing, cost and performance, and deployment work. |
| Machine Learning | Basic workflows, features, training, tuning, evaluation, deployment. | Distributed training, advanced MLflow, MLOps, retraining, drift, rollout. |
| Length | Data Engineer 45 questions in 90 minutes. ML 48 in 90 minutes. | Both: 59 questions in 120 minutes. |
| Best preparation | Do the tasks yourself in a workspace. | Operate real systems, debug failures, and make trade-offs. |
7. Cost, validity, retakes, and renewal
Cost
Every exam page shows a USD 200 registration fee. Taxes can apply, and checkout may show your local currency. A retake is a new attempt at the full fee, and Databricks does not issue free retake vouchers. Two exams cost USD 400 before taxes, so plan the order on purpose.
Validity and renewal
A certification is valid for two years from the date you earn it. To renew, you pass the current version of the exam. The exam guide for Data Engineer Associate mentions a Renewal Exam as an option, and the catalog has a Renewals and Bridge Certifications tab. The certification FAQ says to take the current live exam. Check both pages when your renewal date comes near, and do not assume a course or badge renews anything.
Retakes
You can retake after 14 days, and the same 14 days apply between every later attempt. Each attempt costs the full fee. Databricks does not publish passing scores. It sets them through statistical analysis and they can change when the exam is updated.
Rescheduling
The FAQ says online exams can be rescheduled more than 24 hours before the start, and test center exams more than 72 hours before. Changes closer than that can cost a fee.
8. Certification, badge, accreditation
Databricks uses three kinds of credential. They are not interchangeable.
Certification
A proctored exam for a role or technology. The eight above each cost USD 200 and last two years.
Use it when you want assessed proof of role skills.
Skill badge
Free, unproctored learning assessments. The catalog lists Databricks Fundamentals, Generative AI Fundamentals, and AI Agent Fundamentals.
Use it when you are new to the platform or a topic.
Accreditation
The catalog lists AWS Platform Architect, Azure Platform Architect, GCP Platform Architect, and Platform Administrator separately from the eight certifications.
Use it when your work is cloud architecture or platform administration.
9. Registration and exam day
- Open the exam page for your certification from the Databricks catalog and read the current exam guide. Databricks asks you to recheck the guide before your exam date.
- Create or sign in to your exam account on the exam delivery platform, Webassessor. The online proctoring is delivered by Kryterion.
- Choose online proctoring or a test center. For online, read Kryterion’s technical requirements and run the system check on the machine and network you will use. A locked-down work computer can cause problems.
- Check the checkout total: currency, taxes, date, time zone, and cancellation rules.
- Prepare your room and ID as the exam provider requires. No test aids are allowed.
- After the exam you see an unofficial result at the end. Databricks reviews the proctor recording for passing attempts and then issues your digital credential, which you can find at credentials.databricks.com.
10. How to prepare
An exam does not replace experience. Use the exam guide as a checklist, do the work in a workspace, and then test whether you can explain your choices. Databricks does not offer practice exams, and it warns that third-party practice results can mislead. Use practice questions to find gaps, not to predict a score.
- Read the live exam guide. Write down each domain and its weight, and mark each topic as familiar, practised, or weak.
- Take the training the guide lists. Databricks also publishes an AI prep guide with advice on studying with a chatbot, and it reminds you to check any answer against the documentation.
- Build one complete project for your role: a pipeline, a dashboard, a model, a RAG app, an agent context setup, or a Spark job.
- Practise the trade-offs. Why this ingestion method, this compute, this control, this deployment approach?
- Use practice questions to find gaps. For each wrong answer, find the concept and repeat it in a workspace.
- Book when you are ready, not when you feel pressure.
An example four-week plan
| Week | Work | Sign you are on track |
|---|---|---|
| 1 | Read the exam guide, review platform basics, finish the introductory training. | You can name each domain and your weakest areas. |
| 2 | Do the core tasks for your role in a workspace. | You can finish the main tasks without instructions. |
| 3 | Break things on purpose: failed jobs, schema changes, bad data, permission errors. | You can diagnose a failure and choose a safe fix. |
| 4 | Revisit weak domains, do timed practice, and check registration requirements. | You can explain your reasoning across all domains. |
Four weeks is an example, not a Databricks recommendation. Adjust it to your experience and the exam.
11. Exams that pair well
You do not need every credential. These pairings make sense when your work spans both areas.
| Pair | Why it fits |
|---|---|
| Data Engineer Associate and Associate Developer for Apache Spark | Platform pipeline skills plus portable Spark DataFrame and tuning knowledge. |
| Data Analyst Associate and Data Engineer Associate | Useful if you move from dashboards into building the tables behind them. |
| Generative AI Engineer Associate and Context Engineer Associate | One covers building LLM apps, the other covers what an agent knows when it answers. |
| Machine Learning Associate and Generative AI Engineer Associate | For ML engineers moving into LLM and RAG work. |
| Data Engineer Associate then Professional | A natural climb once you run production workloads. |
12. Mistakes to avoid
- Choosing by job title. A data engineer and a data analyst can both write SQL, and the exams test different things.
- Assuming Professional is automatically better. Pick the exam that matches work you can show.
- Assuming Associate is required first. It is not. It is a sensible step.
- Studying from an old outline. Exam guides change. The current ones are dated August 2026.
- Treating a badge as a certification. They are different credentials.
- Forgetting renewal. Plan for the two-year renewal and its fee.
- Trusting pass-score claims. Databricks publishes no passing score, so nobody else can promise one.
- Expecting a job or a raise from a badge alone. Employers look at projects, SQL and Python skills, and design judgement too.
13. Frequently asked questions
How many Databricks certifications are there?
Eight proctored certifications are in the catalog: Data Analyst Associate, Data Engineer Associate and Professional, Machine Learning Associate and Professional, Generative AI Engineer Associate, Context Engineer Associate, and Associate Developer for Apache Spark. Skill badges and accreditations are separate.
How much does a Databricks certification cost?
USD 200 per exam according to every exam page, plus taxes where they apply. A retake costs the full fee again.
Which Databricks certification should a beginner take?
Choose by role. Data Engineer Associate suits people building pipelines, Data Analyst Associate suits Databricks SQL work, Machine Learning Associate suits foundational ML, and Associate Developer for Apache Spark suits Python DataFrame skills. The free Databricks Fundamentals badge is a good first step on the platform.
Do I need the Associate exam before the Professional exam?
No. Databricks lists no formal prerequisite for the Professional exams. Associate is a sensible step, but Professional exams expect real production experience.
How long does a Databricks certification last?
Two years. To renew, you pass the current version of the exam.
What is the passing score?
Databricks does not publish passing scores. They are set through statistical analysis and can change as exams are updated.
How long do I have to wait to retake an exam?
14 days. The same wait applies between every attempt, and each attempt costs the full fee.
Can I take Databricks exams online?
Yes. Every exam is offered online with a proctor or at a test center. Run the Kryterion system check before exam day.
Are Delta Sharing and Lakehouse Federation on the Data Engineer Associate exam?
Not in the current outline. Data Sharing and Federation is a domain on the Data Engineer Professional exam.
Are badges the same as certifications?
No. Badges are free, unproctored learning assessments. Certifications are paid, proctored exams. Accreditations are a third category in the catalog.
Will a certification get me a job?
It can show product knowledge, but it does not guarantee a job. Hiring also depends on projects, SQL and Python skills, design judgement, and what the employer needs.
14. Keep going on Ultimate Info Guide
15. Official sources
Names, question counts, fees, languages, and rules change. These pages are the source of truth. This guide was checked on 7 November 2026.
- Databricks certification catalog: certifications, accreditations, and badges.
- Certification FAQ: retakes, rescheduling, results, and renewal.
- Data Analyst Associate
- Data Engineer Associate
- Data Engineer Professional
- Machine Learning Associate
- Machine Learning Professional
- Generative AI Engineer Associate
- Context Engineer Associate
- Associate Developer for Apache Spark
- Databricks exam registration
