Class track
Production ML on GCP
A systems-focused course on getting ML from notebook to reliable production.
1. Pipeline architecture
Vertex AI pipelines, components, artifacts, metadata.
2. Feature engineering
Offline features, point-in-time correctness, reuse.
3. Serving
Online and batch prediction patterns.
4. CI/CD
Testing, promotion, environments, and reproducibility.
5. Observability
Model, data, system, and pipeline monitoring.
6. Cost
Compute, serving, storage, and operational tradeoffs.
7. Security
IAM, service identity, secrets, private networking.
8. Platform design
Standardization without blocking teams.