Yesterday's pipeline-created models must be systematically stored somewhere and pass someone's approval before going to production. If we cannot answer "which data trained this model, what is its performance, who approved it, and which version is deployed now?", that is not MLOps—just a pile of files. SageMaker Model Registry is a model catalog and governance tool to answer those questions.
Today we cover Model Registry's essentials: Model Package Groups, model version management, and approval status workflows. MLA-C01 repeatedly asks about Model Registry in scenarios involving "tracking model versions, passing approval, connecting to deployment".