AWS Data Exchange
AWS Data Exchange is a managed service for sharing and managing data entitlements from other organizations. Data receivers can accept private grants or subscribe to third-party data products in AWS Marketplace, then use the data with AWS analytics and machine-learning services.
AI Practitioner focus
- Use it to obtain governed external/third-party data for an AI/ML project; it does not clean, label, train, or analyze the data.
- Supports file, API, Amazon Redshift, Amazon S3, and AWS Lake Formation data sets.
- A data grant gives a specific AWS account access to a dataset; an AWS Marketplace data product packages one or more datasets for subscribers.
- Evaluate licensing, consent, privacy, provenance, representativeness, allowed uses, update cadence, and regional constraints before using external data for training or RAG.
- Common flow: AWS Data Exchange -> Amazon S3 or direct dataset access -> AWS Glue/analytics -> Amazon SageMaker AI or Amazon Bedrock.