Which lifecycle stage involves decommissioning models and data?

• A. Deployment
• B. Retirement
• C. Monitoring
• D. Problem definition

Answer: (B)

The concept being tested is end-of-life management for models and data. In the retirement stage, you plan and execute the decommissioning of models and the disposal or archiving of data. This includes shutting down production endpoints, ensuring data is deleted or anonymized according to retention and privacy policies, and preserving artifacts (like selected training data, model weights, and documentation) for auditability if needed. Retirement is where the lifecycle explicitly covers removing components from active use and handling data responsibly after they are no longer needed.

This matters because other stages focus on how to build, deploy, and monitor models, not how to terminate them. Deployment brings a model into production; monitoring checks ongoing performance; problem definition scopes the work. None of those stages center on decommissioning or data disposal the way retirement does.