Databricks Data Engineer Professional

It is a proctored, multiple-choice certification with 59 scored questions and a 120-minute time limit. No test aides are allowed. It can be delivered online or at a test centre, and is offered in English, Japanese, Portuguese (BR), and Korean. Exams may also contain unscored items, and additional time is factored in to account for that content.

The official certification page reproduced here does not state a passing score or scaled score requirement, so treat any specific number you see elsewhere with caution. Prepare to a high standard across all ten domains rather than targeting a threshold, since the largest domain alone accounts for 22% of the content.

Developing Code for Data Processing using Python and SQL is the largest at 22%, then Cost & Performance Optimisation at 13%. Data Transformation, Cleansing, and Quality, Monitoring and Alerting, Ensuring Data Security and Compliance, and Debugging and Deploying are each 10%. Data Ingestion & Acquisition and Data Governance are 7% each, Data Modelling 6%, and Data Sharing and Federation 5%.

There are no formal prerequisites, but the source recommends hands-on experience performing the data engineering tasks outlined in the exam guide, and says related training is highly recommended. In practice you should be comfortable writing production Python and SQL against Delta Lake, Unity Catalog, Auto Loader, declarative pipelines, and Lakeflow Jobs before sitting the exam.

Follow the sequence the certification page recommends: read the Data Engineer Professional Exam Guide, take the related training it outlines, use the AI Prep Guide, register for the exam, check the technical requirements and run a system check, then review the exam guide again to identify gaps, study to close them, and sit the exam.

Registration is $200. The certification is valid for two years, and recertification is required every two years to maintain certified status. To recertify you must take the current version of the exam, so it is worth reviewing the exam guide again before a recertification attempt.

The certification page reproduced here does not describe a retake policy, including any waiting period or fee for a second attempt. Check the exam delivery platform and the official certification FAQ for the current retake terms before you book, and plan your first attempt as though a retake may involve additional cost and delay.

Two patterns stand out from the domain weights. First, under-preparing for performance and cost work, which is 13% and demands real tuning judgement rather than recall. Second, treating the lower-weighted domains as optional: Data Modelling, Data Ingestion, and Data Sharing together still account for 18%, and the exam expects end-to-end pipeline reasoning across all of them.

Code examples in the exam are primarily in Python and SQL. Beyond that, the exam tests Delta Lake, Unity Catalog, Auto Loader, Apache Spark Declarative Pipelines, Databricks Compute including serverless, Lakeflow Jobs, and the Medallion Architecture, plus deployment with the Databricks CLI, REST API, and Asset Bundles.