AI Agent Hackathon Episode 6: Cognition
The next frontier is an open problem.
What can AI agents contribute to mathematical discovery—and how do we make that progress useful, reproducible and verifiable?
Episode 6 brings together builders, students, mathematicians and researchers to explore two directions: using AI agents to tackle unsolved mathematical problems, and building tools that help with the research process.
You might investigate an open problem, test a conjecture, search for counterexamples, formalise a proof step, or build tools for finding relevant literature and supporting research workflows. The aim is to produce work that others can examine, reproduce and build on. A useful research tool or a carefully verified step forward can matter alongside a complete solution.
The programme
22 November — Opening ceremony, talks and panels
23–27 November — Workshops and project development
28 November — Demo day
After demo day — Judging; timings to be announced
Sponsors & partners
Title sponsor: Cognition
Partners: Modal, Conjectures.io, Bittensor, Cloudflare and DPT
What to expect
A week of building, research and exchanging ideas, with workshops to support participants and a demo day to share your methods and results. Bring a problem you care about, an engineering skill, or the curiosity to learn with a team.
Venue, exact session times, speakers, challenge details and prize information are coming soon.
Beyond the hackathon, our next House of Lords community meetup is planned for February or March 2027. Further details and registration for that meetup will be announced separately.
Bring a question. Build an approach. Show what others can verify and build on.