Inspiration
The inspiration for Partners came from a personal “failure” at the Paris Innov'Hack 2026. I showed up to the event solo, full of energy but without a team. I spent the first few hours awkwardly networking, trying to find people whose skills actually complemented mine.
I realized then that while hackathons are about building, the most stressful part happens before the clock even starts: finding your people. I told my friend:
“I wish there was a matchmaking app that didn’t just list skills, but actually proved we could build together.”
That conversation was the birth of Partners.
What It Does
Partners is a high-reasoning Synergy Engine. It allows builders to instantly generate professional profiles by analyzing their public GitHub repositories—translating raw code into a clear technical identity.
It then uses Gemini 3 to explain the “Why” behind a match through Cognitive Reasoning Traces. Whether you're looking for a co-founder or just a partner for a “fun” weekend project, Partners removes the social friction and gets you straight to the code.
How We Built It
- Backend: FastAPI for handling high-concurrency requests
- AI: Gemini 3 Pro via the Google GenAI SDK
- Reasoning: Leveraged the model’s new Thinking configuration to generate deep-dive synergy reports
- Frontend: Responsive React application
- Design: “Hacker-Dark” Tailwind aesthetic, optimized for fast-paced hackathon use
The goal was to make the experience as fast and focused as a 10-hour Paris hackathon sprint.
Challenges We Ran Into
One of the biggest challenges was handling rate limits during live AI analysis while users onboarded rapidly. I also had to design a system capable of accurately reading a GitHub profile without being overwhelmed by data noise.
Finally, building the project while battling a sore throat on the final day of the hackathon forced me to innovate—I used AI voice synthesis for the demo video itself, making the project even more meta.
Accomplishments That We’re Proud Of
I’m proud of creating a tool that I would genuinely use. The GitHub Bio Generator works remarkably well, turning a list of repositories into a compelling technical narrative.
Most of all, I’m proud of transforming a frustrating experience in Paris into a functional tool that can help the next solo developer find their dream team.
What We Learned
I learned that Agentic AI is the future of networking. By letting AI handle the vetting and summarization, we can focus on real human connection.
I also learned the importance of grounding AI responses in real-world data—like live 2026 hackathon schedules—to make suggestions truly actionable.
What’s Next for Partners
Next steps include:
- Real-Time Chat
- An Instant Team feature where solo developers can join event-specific lobbies
- Support for GitLab and Bitbucket to expand the builder ecosystem
Partners is just getting started.
Built With
- css**
- fastapi**
- gemini
- grounding
- high
- level
- lucide-react**
- project-histories
- python**
- react**
- tailwind
- thinking
- typescript**
- uvicorn**
- vite**
Log in or sign up for Devpost to join the conversation.