Your question is Design ML Lineage and Versioning. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
You're setting up ML pipelines and want a clear way to track how datasets, models, and training runs relate to each other over time. You also need a versioning approach that supports reproducibility, rollback, and auditability when inputs or pipeline definitions change.
Design a lineage tracking and versioning system for datasets, models, and training pipelines.