A weekly model performance report combining CloudWatch metrics, live evaluation, and the
MLflow baseline from project 5 (a --dry-run mode works without AWS credentials). A model
nobody is watching is a model nobody can defend in the next incident review — this project
is the artifact you point to when someone asks “is it still working?”
Translating CloudWatch metrics and live evaluation into a weekly report is exactly the communication skill that separates a senior ML Engineer from someone who can only train models in isolation.