Inspiration Campus life is scattered across portals, chats, and flyers, so students miss events, resources, and each other. We wanted one Rutgers-first place that feels like a real “campus OS,” powered by AI students already trust.

What it does Rutgers Smart Campus Connect AI brings FAQs, personalized suggestions, events, navigation, buzz-style activity views, a virtual game-room queue, and LFG into one Streamlit dashboard. Registrations and queue joins persist in SQLite so lists and counts stay real as people use the app.

How we built it We used Python and Streamlit for the app, Google Gemini for the assistant, scikit-learn for lightweight personalization, and SQLite for events/participants and queue data, plus Rutgers-themed UI and optional Java/HTML pieces for depth. We wired env-based API keys, model fallbacks, and a small test script so Gemini stays reliable across API changes.

Challenges we ran into Gemini model IDs and sidebar styling fought us—404s on older model names and global CSS that hid selectbox text until we scoped styles properly. Streamlit’s layout limits also meant we leaned on CSS and clear module navigation instead of a full native mobile shell.

Accomplishments that we're proud of We shipped an end-to-end story judges can demo in minutes: AI that’s central, not decorative, plus persistence that proves the “platform” idea. The Rutgers identity reads immediately—scarlet, campuses, shuttles—while still feeling like a product, not a slide deck.

What we learned Small, honest integrations beat fake “full stack” claims: one solid AI flow plus real saved state is more convincing than many shallow tabs. Environment setup, API drift, and UX contrast in dark sidebars are first-class hackathon risks worth time-boxing up front.

What’s next for Rutgers University – Smart Campus Connect AI Next we’d plug in real SSO, campus APIs (LMS, transit, room booking), and push notifications so queues and events update live. After that, admin analytics and optional RAG over official Rutgers docs would turn it from a demo into something departments could actually pilot.

Share this project:

Updates