Set up a project

Step 0: choose the mode#

The mode determines which panel reads the field, and it is chosen when you create the project.

Pick the one that matches what you are actually selecting for. A hackathon judged on the pitch rubric will reward the team that demoed well over the team that shipped.

Step 1: details#

The event itself: name, what the program is, and the window you are working to. This is the information that appears to teams on a public submission page, so write it the way an applicant will read it.

Step 2: criteria and weights#

The default pitch rubric arrives with six weighted dimensions and anchor descriptions per band. Two things to decide:

• Do the default weights match your program? Market attractiveness and Team / founder fit carry 0.20 each; the other four carry 0.15. A demo-day cohort that has had equal coaching on story often shifts weight toward Team (P5) and Feasibility (P6) and away from pitch polish.
• When do they lock? Weights are editable right up until the run starts, then they freeze, so the whole field is ranked on one standard end to end.

Details and the anchor table: Criteria and weights.

Step 3: judges#

Configure the panel for this project. In the current release, the judge lenses and their routing — which judge influences which dimension, and how much — are the fixed method of the mode you chose; what you control at project level is the criteria and their weights. See The six judges for the routing matrix.

Step 4: how decks are collected#

Two ways to fill the batch, and you can use both in the same project:

• Add teams yourself. From the project page: name, project, pitch deck, optional note for the jury.
• Open a submission page. Publish a public page, share one link, and teams upload their own decks after signing in with Google.

Access rules, the submission window, invites and sharing are covered in Collect submissions.

Step 5: final review#

The last screen is the one worth slowing down on, because two things become hard to change afterwards: the weights lock when scoring starts, and a published submission page has already been shared.

A short pre-flight:

• Do the weights reflect what your jury will actually argue about?
• Does the submission page say what teams must send, and by when?
• Does your program's rules document have a methodology statement in it — what the AI panel assists with, what the judges decide, and how a team can ask about its own record?
• Do your judges know they will receive a briefing rather than a blank scorecard?

What happens after setup#

The project appears on your dashboard with its status. Nothing is evaluated until you launch a run, and a run needs at least one entry marked ready.

Next steps#

• Collect submissions — publish the page or add teams by hand.
• Brief your jury — the pack to send a week before, and why it beats feedback after.
• Run an evaluation — readiness checks and what each pipeline stage does.