Yesterday we built pipelines (Day 1) and Model Registry (Day 2). But if every code change requires manual pipeline re-run in a notebook, manual approval clicks, and manual endpoint updates, it is still manual operations. CI/CD (Continuous Integration / Continuous Deployment) automates the chain from code commit to model deployment. Bringing software engineering's CI/CD directly to ML is the final MLOps puzzle.
Today we cover AWS ML CI/CD tools: SageMaker Projects and the supporting CodePipeline / CodeBuild / CodeCommit integration plus automated deployment flow. MLA-C01 tests this combination in scenarios: "make code changes auto-trigger model retraining and deployment".