MLREL-06: Enable CI/CD/CT automation with traceability - Machine Learning Lens

MLREL-06: Enable CI/CD/CT automation with traceability

Enable source code, data, and artifact version control of ML workloads to enable roll back to a specific version. Incorporate continuous integration (CI), continuous delivery (CD), and continuous training (CT) practices to ML workload operations. This will enable automation with added traceability.

Implementation plan

  • Use Amazon SageMaker Pipelines- Manual changes to a system can cost additional time and impair reproducibility. Changes to an ML workload should be conducted, tracked and rolled back automatically. MLOps is a collection of best practices around integrating and deploying reproducible, auditable changes. MLOps increases your productivity while automating all facets of your ML development cycle (MLDC). Amazon SageMaker Pipelines is the first purpose-built, continuous integration (CI), continuous delivery (CD), and continuous training (CT) service. With SageMaker Pipelines, create, automate, and manage end-to-end ML workflows at scale.

Documents

Blogs

Videos

Examples