A privacy-first Android app that analyzes toddler walking videos entirely on-device: it checks capture quality, runs 33-landmark pose estimation locally, applies deterministic gait analysis, and has a local AI assistant explain the results to parents (a research prototype, not a diagnostic tool).
The constraint and the product value point the same way here, which is rare. Video of a small child is the last data a parent wants leaving the house, so on-device inference - the theme of the hackathon itself - is not a flourish but the reason the app is usable at all. The second strong choice is where the line is drawn: gait analysis is deterministic, and the language model only explains results. Calling it a research prototype rather than a diagnostic tool shows restraint.
Some winners chose to assume no network and no cloud — and won precisely because of that handicap.
Good intentions do not win. Every public-good project in this archive nailed the operational details.
Community-run events and platform-hosted events ask for fundamentally different things.
Projects that explicitly refuse to let a generative model decide tend to score higher, not lower.
Summaries are written by this site. Project pages include demo videos, screenshots and the team's own write-up (videos may autoplay).