Dates

Event: Friday, October 30, 2026, 12:00 PM to 6:00 PM Central Time

Check-in and lunch: 12:00 to 12:45 PM

Hacking: 1:15 to 4:00 PM

Submission deadline: 4:30 PM Central Time

Demos and judging: 4:30 to 5:30 PM

Winners announced: 5:30 PM at the awards ceremony

Eligibility

Open to students of any school and major, at any skill level. Participants must register on OrganizerHQ, check in in person at JTS (STC Lab) on October 30, and agree to follow the MLH Code of Conduct. Teams have 1 to 4 members. Organizers, judges and their immediate family members may not compete.

Project and Submission Requirements

Build an original project at the Hack Day. Work specific to your project should start at the event. You may use existing open-source libraries and tools.

Your project must:

Use open-source or open-weight AI as an important part of how it works

Be published in a public GitHub repository with an open-source license (for example MIT, Apache-2.0 or GPL-3.0)

Work well enough to demo

Agent skills must follow the Agent Skills Open Standard. Model harness projects must be original work or meaningful changes to an existing open-source harness.

To submit, one team member submits by 4:30 PM Central with: project name, short description, link to the public GitHub repository, a demo link or short video, and the technologies used. Select every challenge you qualify for.

Best Open-Source AI Project with Snowflake: your project must also use Snowflake CoCo and a freely accessible Snowflake dataset. Name the dataset and show how CoCo helped.

Prizes

Best Open-Source AI Project: an MLH DEV Badge for every member of the winning team, MLH belt bags (4 per team), and local prizes announced at the opening.

Best Open-Source AI Project with Snowflake: Arduino Tiny Machine Learning Kits for the winning team.

Prize availability can change. We will confirm the final list at the opening session. Every attendee who checks in receives free swag from MLH while supplies last.

Judging Criteria and Winner Selection

Judges first check that each project meets the requirements. Eligible projects are then scored from 1 to 5 on:

Technical execution

Creativity and originality

Use of open-source AI

Presentation and clarity of what the team built

Winners are chosen from the judges’ combined scores. Judges may discuss close results, and their decisions are final. A team may win more than one challenge if it meets each challenge’s requirements. Winners are announced at the awards ceremony.

Code of Conduct (Devpost requires one for in-person events)

All participants, organizers, judges and volunteers must follow the MLH Code of Conduct: https://mlh.io/code-of-conduct