The Great RAG-athon
Build a RAG system that actually finds the right answer.
A weekend hackathon for anyone who wants to build a real retrieval-augmented generation system — pick your own dataset, wire up retrieval, and show us it actually beats a plain LLM call.
Rules
- Teams of up to 4 (solo is fine too).
- Build a working RAG-based retrieval/QA system over a dataset of your choice.
- Submit a working demo and your source code by the submission deadline — no swaps after judging starts.
- Judged on: does it work, code quality, and how much retrieval actually improves answers over a plain LLM call.
- Standard code of conduct applies — be respectful, no plagiarized submissions, credit any external libraries or datasets you use.
Prizes
₹30,000 cash-equivalent prize pool for the winning team(s).