AWS Certified AI Practitioner (AIF-C01) · Objective 3.3
Retiring a model safely
3.3Standard
A legacy model has an end date.
Migrate without breaking users or audit.
Choose two practices.
Hint 1: Focus on the requirement
Inventory dependencies and users, migrate in stages with comparison and rollback, then retire old data and access deliberately.
Hint 2: Key terms
Review model retirement planning.
Model retirement
The controlled removal of a model version from active service.
Migration plan
A sequence for moving users, data, configuration, and traffic to a replacement.
Dependency inventory
A list of systems, prompts, indexes, data, and users relying on an asset.
Decommissioning
Removing an old resource, access path, or artifact after approved transition.
× Incorrect
Inventory dependencies, versions, data, prompts, evaluations, and users, then plan comparison, staged switch, notice, rollback, retention, and deletion.
Detailed explanation
A
IncorrectDelete old logs and settings immediately
Audit and recovery are lost.
B
CorrectPlan old-versus-new comparison, staged switch, notice, rollback, and retention or deletion
Retirement is controlled.
C
IncorrectWait until shutdown to find a replacement
Service interruption results.
D
CorrectInventory services, versions, data, prompts, evaluations, and users and set a migration deadline
Impact is known.
Check official documentation
Compare the following points with the official documentation.
Path & Pass is an independent learning site and is not affiliated with any certification body. Questions are original practice content based on published objectives.