The whole setup is built to be fair, and to reward real work over polish. Two things decide who advances: who evaluates at each gate, and the five-criteria rubric they use. The criteria are equal and constant across fall and spring. What rises is the evidence bar: everything rests on demonstrated, working evidence, not promises. A polished concept with nothing behind it won't score.
Each group works at a different point in the funnel. And we keep support separate from judgment on purpose: reviewers and mentors never double as the judges who decide who advances. That way the build room stays a place you can actually ask for help.
Score the entry deliverable against a light gate and advance the semi-finalists. A read on the idea and the team, not a product, since nothing is built yet.
A standing panel that follows teams across the build sprint, reading each team's project log and holding two short check-ins, then selects the top 10. Interested observers on the other Fridays, not graders.
A mostly external, senior panel runs the interactive demos and selects the finalists. The deepest, most hands-on evaluation of the fall, decided on working evidence.
Nothing is built yet at entry, so the first gate is lighter than the rubric below. Reading the one-page PDF (and the optional visuals page) and the two-minute video, reviewers look for a clear finance problem, a defined user, a solution that can realistically be built, and feasibility within the timeline.
Each criterion is worth 20% and applies in both phases. The set stays constant; what changes is how each one is read for what that phase actually tests, as the evidence bar rises from a working prototype in the fall to a refined product in the spring.
| Criterion (20% each) | Fall build phase | Spring showcase |
|---|---|---|
| Aesthetics / UI20% | The interface is clear and reasonably polished for a working prototype. | A polished, professional product and presentation. |
| Ease of use20% | A real user can complete the core task in the demo without hand-holding. | Demonstrated usability, with evidence of real use or testing. |
| Finance understanding20% | Sound finance and domain reasoning: a real problem, credible assumptions and data. | Depth of finance and domain reasoning, and validated claims. |
| AI implementation20% | Effective, appropriate, and disclosed use of AI and vibe-coding to build the solution. | Sophisticated, responsible AI use, clearly disclosed. |
| Creativity / uniqueness20% | Original problem framing or approach, differentiated from off-the-shelf tools. | Originality, venture potential, and differentiation. |
Registration for the fall build phase is open. Bring a finance problem worth solving, and we'll help you build the rest.
Register your team