Inspiration
Within the first two weeks of our college experience, our new bright youthful minds were excited to branch out within and outside of our major to find a community we could really bond with. With sparkles in our eyes, we jumped to Instagram: following up club after club, bookmarking website after website, and an hour later, we were content with a list of clubs. By the time we started some of these applications/recruitment processes, we realized that many of the times collide and there are many extraneous pieces of information scattered throughout the respective instagrams. After making many revisits to these Instagram pages and noting down links, dates, and locations, we realized this was not the way to go and that it had to change.
Hence, Me&UDub was born: an accessible, student-centered platform designed to make discovering clubs as easy and engaging as swiping through your favorite social app. Inspired by Tinder's familiar swiping interface, Me&UDub brings personalized club recommendations, searchable organizations, and essential recruitment information together in one place. Because finding your community in college shouldn't feel like another assignment.
What it does
Allows you to create your own profiles of what you're looking for in a club and traits of yourself that may be applicable. According to your profile, Club Tinder shows relevant RSOs at UW where you can then swipe left or right according to if you like it or not.
How we built it
The heart of Me&UDub is its matching system, powered with Anthropic API. After the user completes their profile, the AI ranks clubs by fits and even reorders clubs as it gets new information from the user. The swiping interface was made with React, and the backend a Python + Fast API app. Club data was sourced from individual club websites, social media, and RSO lists, and is hosted in a SQLite database.
Challenges we ran into
Getting accurate data was a major challenge we ran into. Clubs didn't always have the same way of publicizing information (hence why it might be hard for someone to join a club), so we had to ensure that our research checked that its information was up-to-date and relevant. Besides data-curation, we also had to find a balance between matching speed and swiping. We do most of the ranking at the start, and re-rank the cards every couple of swipes in the background with a lighter model so the app feels reactive, but not slowed down.
Accomplishments that we're proud of
We're proud of making a user-friendly interface and a relatively fast app (what we think), for our first time using Anthropic API. We're also proud of hosting a project on a custom domain for the first time, and doing it through Google Caddy.
What we learned
It's hard to get an idea in the first place, and you won't even know you don't want to continue with an idea until you start building. We went through many iterations of what we wanted to make until we settled on something we felt we could be proud of.
What's next for Me&UDub
The future of Me&UDub is the data that drives it: expanding the list of RSOs and options to get supplement data directly from an organization are both ideas for future add-ons.
Log in or sign up for Devpost to join the conversation.