Module 9: Serving and Inference
1 min read·
What you need first
- The warning threshold and the logistic baseline (Module 3 lesson 3), training-serving skew (Module 5 lesson 4), latency percentiles (Module 8 lesson 3) and what an HTTP request is. The lab's Step 1 rebuilds Modules 1 to 8.
- No model calls and no keys. Docker and Compose files are written here, not run.
- If you already build FastAPI services with admission control, timeouts and fallbacks, go to the lab.