ML Artifacts Ownership Enforcement
summary
Centralized inventory enforces ownership for every ML artifact by registering models, features, pipelines and code in an entity framework backed by a graph database inventory that records DAGs and owner attributes. A one-time backfill populates missing owners across legacy platforms and thousands of features, then validators block creation of new artifacts without ownership and an ownership observer updates owners automatically to managers when people leave, which reroutes alerts correctly, enables privacy review, generates metrics and saves on-call triage time.
AI-written summary · about
chapters (8)
- 0:00Intro
- 1:42Story
- 5:22Problems
- 6:49Solution Workflow
- 8:53Solution Architecture
- 13:42Demo
- 15:21Future Work
- 17:11Conclusion
related talks
32mMachine Learning Models Versioning Using Open Source Tools
33mBuilding Resilient (ML) Pipelines for MLOps
37mMetaOps: Metadata Operations For End-To-End Data & Machine Learning Platforms
27mEmpowering Traceable & Auditable ML in Production
12mIAM and Privileged Access Foundations for Secure ML Platforms
34mScaling ML in the Traditional Enterprise
50mBuilding AI Security In: MLSecOps in Practice
29mML in Production on Databricks