Skip to content

Lab 09: CI/CD Regression Gate

Series: AI Eval Labs · Phase: Production Pipelines · Depends on: Lab 01, Lab 02

Core scope

  • [ ] GitHub Action running parameterized evals on every PR
  • [ ] Block merge if task success drops 2% (significance-checked) or latency spikes
  • [ ] Publish eval summary as a PR comment

Stretch

  • [ ] Cached baselines and per-tag gating