MLOps

Canary Deployment

Ramping a new version through 1% → 5% → 25% → 100% with automated guardrail thresholds.

Why interviewers ask about it

Automate the kill thresholds so rollback needs no human, and make rollback a config flag against a warm artefact - not a redeploy.

Running a new model on real traffic without serving its output, comparing scores and latency against the incumbent.

This term is part of the free AI/ML Engineer interview preparation module - browse the full glossary for every definition.