Microsoft · AI-300
Foundation models in production
AI-300 · Lesson 10 of 15 · 1:07:53
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: serverless deployments and managed compute, Global Standard, Data Zone, and Batch, choosing a model, version upgrade options, and provisioned throughput with spillover. Includes three original practice questions.
Preview the topics
- All inference processing must stay inside the European Union, and be billed per token. Which deployment type fits?
- A production deployment must never change model version until the team’s own tests pass, even if that means planning upgrades before retirement. Which version upgrade option fits?
- A provisioned deployment handles the normal load, but Monday peaks cause 429 errors. The overflow must be served without buying more PTUs. What should the team configure?