Microsoft · AI-300
Registration, versioning, and responsible AI
AI-300 · Lesson 6 of 15 · 37:01
YouTube loads only after you press play. This chapter opens at its timestamp in the full course. Watch on YouTube ↗
In this lesson
Prepare for the AI-300 exam (Microsoft Certified: Machine Learning Operations Engineer Associate). In this lesson: MLflow models and the MLmodel file, registering from a run, versions and archiving, the feature retrieval specification packaged with a model, and the Responsible AI dashboard. Includes three original practice questions.
Preview the topics
- A model version was replaced in production, but two batch jobs still reference it. The team wants it hidden from the model list, without breaking those jobs. What should it do?
- A model trained with features from a feature store must look up exactly the same features at inference time. What should be packaged with the model artifact?
- A loan model is accurate overall, but the risk team suspects that it fails mostly for one group of applicants. Which part of the Responsible AI dashboard shows where the errors concentrate?