uOGenAI
The University of Ottawa’s generative-AI and retrieval platform, built as lead developer and co-presented at the Ottawa Applied Medical AI Summit in May 2026.
Overview
Designed its initial architecture and implemented its vector search for retrieval-augmented generation, taking it from prototype to pilot.
Built on a NestJS backend (Prisma, PostgreSQL) with LangGraph agents on AWS Bedrock, and approved for AWS with Canada-only hosting.
A single access point to several language models, with question answering grounded in uploaded documents and custom assistants that teams can share.
Designed for privacy and governance from the start and priced per use rather than per seat. In a pilot with about 300 testers since March 2026.
As the platform’s lead developer, co-presented ‘uOGenAI: A Platform for Retrieval-Augmented Generation at uOttawa’ with Salim Manseur at the summit hosted by OMARI and the CHEO Research Institute, including a live demo.
Tools & technologies
- RAG
- LLM platforms
- Document Q&A
- Governance
- Privacy
- System integration
- NestJS
- LangGraph
- AWS Bedrock
Photos3