SageMaker Model Monitor: data capture on the live endpoint, a baseline from training data, deliberately injected drift, and the resulting violation reports. This is the project that proves the model from project 7 will not just degrade silently — and it directly feeds the retraining trigger in the next project.
Hands-on experience with SageMaker Model Monitor — baseline, data capture, drift injection, violation reports — is a named, screenable skill on most ML Engineer and MLOps job descriptions.