MLOps checklist components

The MLOps checklist consists of nine different components. Each component represents an aspect of a mature MLOps system. This section covers the checklist components in detail. Note that there is some overlap in this section because certain components depend on other components.

The checklist covers the following areas:

• 1. Data-centric management

• 2. Experimentation

• 3. Observability and model management

• 4. Robust pipelines and promotion

• 5. Continuous integration

• 6. Continuous monitoring

• 7. Continuous deployment

• 8. Continuous training

• 9. Governance

Note: Security is already well-covered by AWS. We recommend looking at security-related AWS Prescriptive Guidance guides and patterns, and reviewing AWS security best practices.