Exam format
dbt Labs’ page sets 65 questions in 2 hours, a passing score of 65% and a fee of $200, on dbt version 1.11. You can take it in English, Japanese, Spanish, French or Portuguese. dbt Labs recommends SQL proficiency and at least six months of work with dbt or dbt Core first. Its exam guide, with sample questions and a study guide, is linked from the same page.
The seven topics run from building models to debugging them, testing, managing packages and sources, and using state to build only what changed. dbt Labs doesn’t weight them, so plan to cover all seven.
Study plan
Our dbt course is a dbt-style IDE in your browser. Every build and test runs, and each task is checked against the models and rows it produces. Work through the units in order: models, sources and tests first, then layering, Jinja and macros, then incremental models, snapshots, packages and contracts. The expert units on debugging and state selection come last because they need everything before them. Do the five projects along the way; the safe model release project practices state selection in CI.
About our mock
The scenario mock follows the exam’s length: 65 questions in 2 hours. The questions are our own, written to the seven topics, and passing it isn’t a dbt certification.
Gaps in our course
The course practices dbt Core commands in a simulated IDE. It doesn’t cover dbt Cloud’s account settings, job scheduler or Semantic Layer, so read dbt Labs’ documentation for anything the exam asks about those. Our IDE isn’t dbt Labs’ product, and The Work Gym isn’t affiliated with dbt Labs.
After the exam
The course has interview sets at three levels, and the Interview Hub has spoken dbt interviews. The analytics engineer pathway puts dbt together with SQL, Git and Snowflake.