CampusLoop just hit 6 more days of public voting, and it already carries an Honorable Mention from CSC Back-to-School (553 participants) — so this post is about the thing I actually want to fix while voting is open.
Every June, dorm hallways fill with boxes of textbooks headed for the landfill. Every September, the same students pay full list price for the same editions. The supply exists. The demand exists. Nobody built the loop.
So I built the loop, and then removed the server.
CampusLoop runs entirely in the browser. Photograph a textbook cover and it identifies the book, values it, and matches it to the student who needs it next — with zero network calls and zero accounts. Turn off your Wi-Fi and the whole product still works. That constraint wasn't aesthetic; it was the whole point. A tool my classmates would actually trust has to work offline on a phone, ask for no account, and keep every photo on the device.
Two things I'd genuinely like feedback on, more than votes:
Fair matching, not a queue. Each round builds a bipartite graph and finds a maximum-cardinality matching with Hopcroft-Karp, then runs a 2-swap fairness pass. My first greedy matcher "looked fine" and quietly starved late joiners. If you have ever had a queue that was technically fair to nobody, this is that bug, fixed.
The classifier shows its confidence AND its 100-dimensional feature vector next to every match. Above a 0.72 margin it auto-confirms; below it a human reviews. I don't want a decision my users can't inspect.
I'm 16, self-taught, and this is my first solo hackathon build. If you have 2 minutes: the 4:47 demo plays right on the page — no signup, no ads. If you have 10: the source is public and MIT licensed, and I would genuinely like to be told what I got wrong.
Log in or sign up for Devpost to join the conversation.