Inspiration
I came to this hackathon alone. A lot of the CS students I know from CT State Community College could not make it today, and that got me thinking. If there was a platform that connected me with those students earlier, people in my major, people with my same interests and goals, I could have had a team. I could have gone even further with this project. That is the real inspiration behind StudyLink. Not a textbook problem, but something I felt firsthand walking into this room today.
What it does
StudyLink is a real-time Android app that helps university students find study partners, groups, and lasting friendships on campus. Students can post a session in 30 seconds, browse and join sessions instantly, message each other directly inside the app, and filter sessions by club or association. Every user builds a social profile with their major, year, courses, clubs, and hobbies so you are not just finding a study partner, you are finding someone you actually connect with.
How we built it
I started by designing the full UI in Figma before writing a single line of code. Wireframes, color palette, component layouts, all planned out first. Then I implemented it in Kotlin using Jetpack Compose for the frontend. Jetpack Compose is a declarative UI framework that let me build every screen without writing any XML, which was a massive time saver working solo. For the backend I used Google Firebase. Firestore gave me a real-time NoSQL database so sessions and messages appear instantly across all devices the moment they are posted. Firebase Auth handled Google Sign-In in about 10 lines of code with no passwords to store or manage. For AI I integrated Gemini 3 Flash through the Generative Language API. I built four live AI features: smart natural language search, personalized session recommendations, automatic session description generation, and a prep checklist that tells you exactly what to study before joining a session. The architecture follows a repository pattern so the UI never talks to Firebase directly. Everything flows through typed repository classes which keeps the codebase clean and easy to scale.
Challenges we ran into
The biggest challenge was getting Gemini working. The model versioning across API endpoints kept changing and I had to debug through several 404 errors before landing on the right endpoint for Gemini 3 Flash Preview on the v1beta API. Once that clicked everything else fell into place. Building real-time chat from scratch in a single session was also a challenge. Getting Firestore snapshot listeners to sync correctly between two users, handling chat ID generation, and making the message bubbles scroll smoothly all required careful attention under time pressure. The hardest part overall was working alone. Every decision, every bug, every design call was mine. But that is also what made finishing it feel worth it.
Accomplishments that we're proud of
Shipping a full working Android app solo in a single hackathon session is something I am genuinely proud of. Not a prototype, not a mockup, but a real app with live data, real-time sync, and four working AI features.I am proud of the design process too. Taking the time to build everything in Figma first before touching code meant the final product looks intentional, not thrown together. The warm sunrise theme, the ocean wave backgrounds, the animated splash screen, all of it was planned before a single line of Kotlin was written.I am proud of getting Gemini 3 AI working live. Smart search, session recommendations, auto descriptions, and prep checklists all powered by real API calls, not mocked data. That took real debugging under real time pressure.Most of all I am proud of the story behind it. I built something that solves a problem I felt personally on the day I needed it most. That connection between the problem and the product is something I will carry into every project I build from here.]
What we learned
Building solo forced me to make fast decisions and own every part of the stack. I learned how to design with intent in Figma before touching code, how real-time listeners work in Firestore, how to integrate a live AI API and handle fallbacks gracefully, and how to ship a full product under pressure. I also learned that the best apps solve problems you have personally felt, not problems you read about.
What's next for Study Link
The foundation is solid and the roadmap is clear. The next step is Microsoft Outlook authentication so students can sign in with their university email address, making StudyLink feel native to campus life from day one. After that the focus shifts to push notifications so students get alerted the moment a session is posted in their course or club. Calendar integration comes next so sessions sync directly to your phone and you never miss one. On the AI side the goal is to deepen the Gemini integration with smarter matching that learns from your study history and suggests partners based on compatibility, not just shared courses. Longer term StudyLink becomes university-specific, with custom onboarding for each school, verified student accounts, and a leaderboard for study streaks and hours logged. The vision is a platform that every CS department, math department, and student organization can adopt as their default study coordination tool. StudyLink started because I walked into a hackathon alone. The goal is to make sure no student ever has to feel that way again.
Log in or sign up for Devpost to join the conversation.