AWS Certified AI Practitioner

Data Lineage, Provenance, and Source Citation

How to document where data came from and what was done to it: cataloging with AWS Glue, SageMaker ML Lineage Tracking, Model Cards, and citations at inference time.

Intermediate 20 minutes 6 Learning Objectives
  1. Distinguish provenance, lineage, and cataloging, and name the question each one answers
  2. Describe what the AWS Glue Data Catalog records and how crawlers populate it
  3. Explain what Amazon SageMaker ML Lineage Tracking captures and why it enables reproducibility
  4. Identify what belongs in a SageMaker Model Card, including intended uses and risk rating
  5. Compare Model Cards, AI Service Cards, and the Model Registry
  6. Explain how Amazon Bedrock Knowledge Bases returns citations with a generated response