Azure DP-100 exam guide covering data preparation, training, deployment, and machine learning operations decisions.
This Microsoft Certified: Azure Data Scientist Associate guide helps DP-100 candidates focus on what the exam tests, where close answers usually split, and which review page to use next.
Use the study plan to group Azure service, identity, and deployment choices, the cheat sheet for scenario decisions, the sample questions for applied practice, the FAQ for scope checks, the resources page for Microsoft Learn exam references, and the glossary when product names blur together.
| Item | Guide value |
|---|---|
| Vendor | Microsoft |
| Exam or credential | Microsoft Certified: Azure Data Scientist Associate |
| Code or shorthand | DP-100 |
| Study level | Associate data science |
| IT Mastery page | DP-100 exam page |
| Guide shape | Start-here page, study plan, cheat sheet, FAQ, resources, and glossary. |
| Lane | What to master | Common weak answer |
|---|---|---|
| MLOps infrastructure | Provision workspaces, compute, registries, environments, identity, networking, and IaC for repeatable ML operations. | Treating notebook success as production readiness. |
| Model lifecycle | Track data, training, evaluation, registration, deployment, monitoring, rollback, and retraining signals. | Ignoring lineage, drift, or deployment approval because the model metric looked good once. |
| GenAIOps infrastructure | Operationalize prompts, agents, retrieval, model endpoints, evaluations, and content-safety controls. | Managing GenAI apps like static code without prompt, retrieval, and model behavior checks. |
| Quality and observability | Measure relevance, groundedness, safety, latency, cost, drift, and incident signals. | Relying on ad hoc manual testing for production AI behavior. |
| Optimization | Tune model choice, prompt design, retrieval, batching, caching, endpoint shape, and cost-performance trade-offs. | Scaling capacity before proving the bottleneck or quality failure. |
If the stem says production AI, answer with repeatability, lineage, evaluation, monitoring, rollback, and ownership.
Use the current Microsoft Learn exam page for live exam details, including name, status, pricing, duration, delivery method, languages, retirement or beta changes, and domain weights where applicable.