5 · Applied & production AI

17. MLOps: Serving, Monitoring and Drift

From notebook to endpoint: packaging, registries, CI/CD, observability and retraining.

10 min read · 3 MCQs

Reproducibility and the registry

Track experiments (parameters, metrics, data version, code commit) and promote artefacts through a model registry with stages: staging, production, archived. A model you cannot rebuild exactly is a liability.

Serving patterns

Batch scoring for offline predictions, real-time endpoints behind autoscaling for user-facing latency, and streaming for event pipelines. Shadow deployments and canary rollouts let you compare a new model against production traffic before switching.

Monitoring and drift

Log inputs, outputs, latency and cost. Data drift means the input distribution moved; concept drift means the input–output relationship moved. Alert on both, plus on prediction distribution shifts, and define a retraining trigger up front.

Chapter quiz

3 questions · pass mark 75%
  1. 1. Concept drift is when…

  2. 2. A shadow deployment…

  3. 3. A model registry provides…

Answer every question to submit. Progress for ai-17 is saved in this browser.