MLflow model registry + shadow deploys #38
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Register every model version in MLflow, with model cards attached. Artifacts in MinIO/S3. Shadow-deploy wiring already exists from #50; this issue focuses on MLflow-integrated promotion: a registered version can be marked shadow/candidate/production and the recommender selects per request via stage label. Gated promotion with logged disagreement (tied to #52 observability).