ConnectX

Inspiration

When Azuka transferred to Texas Tech, he spent his first week on campus trying to find his people. What he found instead was a scavenger hunt. Org listings were scattered, most were outdated, and the only real way to know if a community fit was to show up to a meeting and find out. Multiply that by dozens of organizations and you get a process that burns out most students before they ever find their place. ConnectX came out of that frustration: finding your community on campus shouldn't be a trial-and-error slog.

What it does

ConnectX turns a short student profile into a personalized map of campus life. Students share their major, classification, interests and hobbies, and optionally their ethnicity, and Gemini cross-references that profile against a live database of student organizations and their events. Instead of a directory dump, students get matches, orgs and events that actually line up with who they are and what they care about.

From there, ConnectX makes it easy to act on those matches. A built-in calendar feature lets students push events straight to their Outlook calendar in one click, and a saved events tab keeps the ones they don't want to miss in a single place.

How we built it

The frontend is React and Next.js, with React styling libraries handling the interface work so the experience stays fast and uncluttered. Behind it, a third-party web scraper runs weekly through an agentic workflow, pulling events from a wide range of campus organizations without any manual upkeep. Everything lands in TigerData, where events are stored and organized under their parent org. Gemini sits on top of that database, reading the current event landscape and reasoning over it to match students to the organizations and events that fit their profile.

Challenges we ran into

Our hardest problem wasn't the matching, it was the data underneath it. Gemini is only as good as what it's reasoning over, so getting a complete, trustworthy picture of campus events was the real work. The complication is that a large share of "active" student organizations aren't actually active. They hold a spot on the official list, post nothing for a semester, and still look alive from the outside. That noise produced false matches and filled the database with entries that wasted a student's time, so we had to figure out how to separate genuinely active orgs from ones that exist on paper.

Accomplishments that we're proud of

The agentic scraping workflow is the piece we're proudest of technically, it runs weekly on its own and keeps the database current without anyone touching it, which is what makes the rest of the product possible.

Beyond the build, we're proud of what this can do for Texas Tech. Campus community shouldn't come down to who you happened to sit next to in a lecture hall. If ConnectX connects even a handful of students to the places, events, and people that make college worth remembering, it was worth building.

What we learned

This project pushed all of us into new territory: integrating the Gemini API and prompting it to reason reliably over real data, standing up and managing a TigerData instance, and using styling libraries to build an interface that feels seamless rather than stitched together. We also got a lot better at using AI to accelerate our own workflow, not just as a feature in the product, but as a way to build faster and iterate more efficiently as a team.

What's next for ConnectX

Scale is the obvious next step. Most universities have no tool like this, and the architecture isn't Texas Tech–specific, point the scraper at a new campus and the system works. Mobile is next on the roadmap, since students are looking for events on their phones, not at a desk. After that, we want to add a chatbot so students can just ask ConnectX a question in plain language and have it pull live data and answer directly.