The short version
The Databricks Certified Machine Learning Professional exam tests whether you can build, operate, and deploy machine learning systems on Databricks at enterprise scale. It has 59 scored multiple-choice questions and a 120-minute time limit, and it splits into Model Development (44%), MLOps (44%), and Model Deployment (12%). This guide follows the exam guide that Databricks marks as the live version as of September 30, 2025.
What the exam is
The credential is called Databricks Certified Machine Learning Professional. As the name says, it sits at the professional level of the Databricks machine learning track, above the associate exam. Databricks describes the person who passes as someone who can do advanced machine learning engineering at enterprise scale, with production-ready systems that include monitoring, testing, and deployment.
That wording is a fair summary of the questions you should expect. They are not “what does this button do” questions. They describe a team, a data size, a constraint, and a goal, then ask which approach fits. Knowing a feature exists is the entry fee. Knowing when to pick it over its neighbor is the real test.
Who it is for
Databricks aims the exam at machine learning engineers and data scientists who already run models on the platform. The recommended background is hands-on: Databricks highly recommends about one year of practical experience with the tasks in the exam guide. In our view, working knowledge of Python and libraries such as scikit-learn, SparkML, and MLflow, and of Lakehouse Monitoring and Databricks Model Serving, will help a lot. No prerequisite is required. The prerequisites page turns that into a self-check.
Format at a glance
| Item | What Databricks says |
|---|---|
| Exam | Databricks Certified Machine Learning Professional |
| Level | Professional |
| Questions | 59 scored multiple-choice questions |
| Time | 120 minutes |
| Passing score | Not published |
| Language | English |
| Delivery | Online proctored or at a test center |
| Test aids | None allowed |
| Validity | 2 years, then recertify by taking the current exam |
The exam may also contain unscored items that Databricks uses for statistics. They are not marked, and extra time is built in for them. Details on scoring, delivery, and retakes are on the exam format page.
The three domains and their weightings
| Domain | Weighting |
|---|---|
| Model Development | 44% |
| MLOps | 44% |
| Model Deployment | 12% |
Model Development
This domain covers SparkML pipelines, estimators and transformers, tuning and evaluation, and scoring in batch or streaming. It also covers scaling and tuning: the pandas Function APIs, distributed hyperparameter tuning with Optuna and with Ray, and vertical versus horizontal scaling. Add advanced MLflow (nested runs, custom metrics and artifacts, custom model objects) and advanced Feature Store concepts such as point-in-time correctness, online tables, and on-demand features.
MLOps
Half of the exam is about running ML as an engineering discipline. Topics include model lifecycle management in a deploy-code setup, unit and integration testing, environment architectures with Databricks Asset Bundles, automated retraining, and drift detection and monitoring with Lakehouse Monitoring.
Model Deployment
The smallest domain, but not an afterthought. It covers deployment strategies such as blue-green and canary, rolling out models with Databricks Model Serving, and custom PyFunc models logged to Unity Catalog and queried through REST or the MLflow Deployments SDK.
Official source
The vendor’s own page is the Databricks certification page for this exam. It and the exam guide it links to are the final authority if anything here differs. Databricks says the guide is updated whenever the exam changes, and suggests checking again about two weeks before your exam date.
More guides for DATABRICKS-MLP
- Databricks ML Professional Study Plan
- Databricks ML Professional Sample Questions
- Is the Databricks ML Professional Exam Hard?
- Databricks ML Professional Renewal and Retakes
- Is Databricks ML Professional Worth It?
- Databricks ML Professional Cheat Sheet
- Databricks ML Associate vs Professional
Next step
Read the domains explained page to see where your time should go. When you want to check yourself, our Databricks ML Professional practice tests are original practice sets built around these three domains.
Frequently asked questions
Does this certification expire?
Yes. Databricks says it is valid for two years from the date it is awarded. To recertify, you take the full exam that is live at that time.
Is there a required prerequisite exam?
No. Databricks lists no required prerequisite, but it highly recommends relevant training and about a year of hands-on experience.
Where do I find the current list of skills?
In the exam guide linked from the Databricks certification page. It names every section and skill bullet, and it is the document to trust over any summary, including this one.