AI knowledge systems
Retrieval-augmented systems built on an organization’s own documents, so people can ask questions of a specialized collection and get answers grounded in it.
- Ingestion: documents, metadata, and full text gathered into a clean, maintained corpus
- Answers that cite the passages they rely on
- Evaluation that shows where the system is right, wrong, or unsure
Current work includes a retrieval-augmented system over a specialized scholarly corpus for a university research center.