Apply the Retrieval-Based AI curriculum to a checked project.
Project brief
Required concepts
- Retrieval corpus contracts: identity, permissions and document lifecycle
- Chunking documents: preserve section identity and answer boundaries
- Retrieval design: compare lexical, semantic and reranked candidates
- Grounded answers: require passage support and a no-answer path
- RAG evaluation: separate retrieval recall from answer support
- Retrieved text as untrusted data: keep instructions and tools separate
- Index freshness: publish complete revisions and remove stale chunks
Completion standard
Submit runnable work, a data or document manifest, measured results and failure-case evidence.