Cerebro

An app designed to help users learn a topic of their choosing, through playing mini-games with friends or random people. It's game pigeon but with an AI powered Q&A feature for embedded learning.

Comment

Inspiration

We were inspired by the growing epidemic of doom-scrolling and its impact on productivity and mental wellness. Working alongside NRVE, we wanted to build something that replaces mindless scrolling with something genuinely valuable. Education shouldn't feel like a chore, and entertainment shouldn't feel like a waste of time. Cerebro is our answer to both.

What it does

Cerebro is a gamified peer-to-peer learning platform that turns everyday mini-games into personalized learning experiences. After signing in, users select topics they want to learn about across professional skills, technical knowledge, life skills, and more. From the home page, players choose from mini-games including Chess, Connect 4, and Tic Tac Toe and select which topic they want to focus on that session. Games are multiplayer so you get matched with a random peer or a friend who is also learning, creating a competitive social environment. During gameplay, every key moment triggers an AI-generated question tailored to that player's topic and learning history. The more you play, the smarter the questions get as Cerebro's adaptive model analyzes your answer history to identify gaps and adjust difficulty in real time. Correct answers earn reward points tracked on a live leaderboard.

How we built it

We used an incremental delivery approach, shipping working features continuously rather than building everything at once. Our stack was React, FastAPI, MongoDB Atlas, Auth0, and the Gemini API. We split the team across frontend and backend in parallel after standing up auth, which let us move fast without blocking each other.

Challenges we ran into

MongoDB Atlas gave us SSL errors due to a Python 3.13 compatibility issue that took time to diagnose. The Gemini API required significant prompt engineering to return consistently structured JSON as it frequently returned malformed responses that broke our parsing logic. We also ran into complexity with the chess game logic around detecting the right game state events to trigger question popups.

Accomplishments that we're proud of

We built a full stack application in 24 hours that tackles a real and widespread problem. The adaptive question engine genuinely adjusts to each user's knowledge level so the learning experience improves the more you play. We are also proud of how well we collaborated under pressure and shipped something we would actually use ourselves.

What we learned

We deepened our practical knowledge across the entire stack including async FastAPI patterns, MongoDB, and prompt engineering with Gemini. Beyond technical skills, this hackathon sharpened our ability to make fast architectural decisions, debug under pressure, and communicate clearly as a team when time is limited.

What's next for Cerebro

  • A mobile app to make Cerebro accessible anywhere and truly replace phone doom-scrolling at its source
  • Expanded game library including reaction-based and physical activity games
  • Richer question formats including voice note responses and open-ended answers graded by AI
  • Social features including friend challenges, guild-based learning, and streak competitions

Built With

Share this project:

Updates