End-to-End Lineage with Amazon SageMaker AI, MLflow and DVC

A team at AWS architected a solution using DVC by lakeFS as part of an end-to-end ML model lineage pattern, alongside Amazon SageMaker AI and SageMaker AI MLflow Apps. Join us to walk through it live with the people who built it!

AWS solutions architects Sandeep Raveesh, Paolo Di Francesco, and Manuwai Korber join us to break down the architecture from the post: how a SageMaker Processing job versions a dataset with DVC, how a SageMaker Training job pulls that exact version and logs it to MLflow, and how that chain lets you answer “which data trained this model” and “can I reproduce it” with just a lookup instead of a multi-day investigation. They’ll also walk through the record-level pattern built for regulated environments to cover the manifests, consent registries, and audit queries that prove a specific record was excluded from training after an opt-out.