ChessHacks is a 36-hour hackathon where participants form teams of up to three people to build an AI model that plays chess. Winners will be determined in a tournament-style format, with bots competing head-to-head. During the event, teams will have the chance to test their bots against others live to see how they compare.

Unlike traditional hackathons that emphasize brainstorming and ideation, ChessHacks is centered around solving one collective problem. Instead of the typical “science fair” style judging ChessHacks uses an objective, tournament-style system. Bots created by participants compete directly against one another, and a single winner is declared based on measurable in-game performance, not subjective preference.

You don't need prior AI/ML experience to participate. However, we encourage participants to have at least a basic understanding—watching 3Blue1Brown's introductory ML series and experimenting with building a simple model is a great start.

We welcome everyone! Priority will be given to students and recent graduates, but anyone with an interest in AI and hacking is encouraged to join.

No, why be intelligent when you have Artificial Intelligence

Cold hard neural networks: no search algorithms, no fine-tuned GPTs, no shortcuts. You must build your own model during the hackathon, whether it's an MLP, CNN, Transformer, or any other architecture you choose. Any public dataset is fair game, and distilling knowledge from chess engines is allowed, but the emphasis is on originality—designing and training your own system within the event.

Participants will receive developer tools for model training, a dashboard to submit and manage bots and view rankings, and several pre-trained models that may be extended. We'll continue iterating on details, but our commitment is to provide everything hackers need to build successfully.

Sponsorship: If you'd like to sponsor, please reach out at [email protected]

Volunteering/Contributing: If you'd like to help organize or contribute in another way, contact us at [email protected]