Level 5: AI / LLM systems · L5.6 Evals & guardrails
RAG metrics and regression evals in CI
Measure faithfulness and retrieval quality and catch regressions.
12 min · Pro
What you take away
- Measure retrieval and generation separately — context recall and precision for one, faithfulness and answer relevance for the other.
And 3 more inside the class, in Pro.
How the class is laid out
- The problem
- The garage measures parts, not goodness
- Running it in CI without a five-figure bill
- Check your understanding