It will describe the data layer for AI products: ingestion, storage, RAG indexing, feature stores, dataset versioning, and boundaries between experimental and production data.
The focus is how the data runtime supports typical portfolio AI products, not an abstract data lake.