Microsoft retired DP-100 Azure Data Scientist Associate on 1 June 2026. AI-300 MLOps Engineer Associate replaces it: what changed, what DP-100 holders should do and how to prepare.
Short answer: Microsoft retired DP-100 (Azure Data Scientist Associate) on 1 June 2026. Its replacement is AI-300, the Machine Learning Operations (MLOps) Engineer Associate. The focus moves from building models in notebooks to running them reliably in production: pipelines, deployment, monitoring and governance.
From data scientist to MLOps engineer
| DP-100 (retired) | AI-300 (current) | |
| Certification | Azure Data Scientist Associate | MLOps Engineer Associate |
| Last exam date | 1 June 2026 | Available now |
| Focus | Training and tuning models in Azure Machine Learning | Taking models and AI apps to production and keeping them healthy |
| Typical skills | Data preparation, training, evaluation, responsible AI | Pipelines, model registry, deployment, CI/CD for ML, monitoring, governance |
The change matches the job market: companies now have more models than ever and need engineers who can operate them. Check the AI-300 skills outline on Microsoft Learn for the exact topics.
I hold DP-100. What happens?
Your Azure Data Scientist Associate certification remains valid until it expires; retired certifications are not renewed. Keep it on your resume. If you now work on deployment and operations, AI-300 is the logical next exam.
What an MLOps engineer does day to day
An MLOps engineer turns a working notebook into a service the business can rely on. That means packaging the training code, versioning data and models, building pipelines that retrain on new data, deploying models behind endpoints, watching for drift and cost, and rolling back safely when something breaks. In Indian companies the role often sits between the data science team and the platform or DevOps team, so people come into it from both sides: data scientists who want to ship, and DevOps engineers who want to work with AI.
How to prepare for AI-300
- Refresh the ML basics: features, training, evaluation.
- Build a pipeline that trains, registers and deploys a model.
- Add CI/CD so a code change retrains and redeploys safely.
- Set up monitoring for drift and performance, with alerts.
- Finish with timed mock exams.
Our Azure MLOps course (AI-300, formerly DP-100) follows these steps on Azure, and MLOps Engineering covers the same practices across tools. See the MLOps engineer roadmap and all exam changes on the certifications page.
Frequently asked questions
Can I still take DP-100?
No. The last day was 1 June 2026.
Is AI-300 only for data scientists?
No. It suits DevOps engineers, ML engineers and data scientists who deploy and run models.
Do I need DP-100 before AI-300?
No. There is no prerequisite, though basic ML knowledge helps.