SkillSwap is an AI-powered peer-to-peer skill exchange platform designed to connect college students who want to teach what they know and learn what they don't. Students across departments often have complementary skills — a CSE student who excels at coding may want to learn graphic design, while a design student may want help with Python. Yet no structured platform exists to bridge this gap on campus. SkillSwap solves this by letting students create profiles with skill tags (what they can teach and what they want to learn), discover matching peers through an AI-driven recommendation engine, and schedule short one-on-one learning sessions. The platform uses the Gemini API to semantically match users beyond keyword overlap — understanding that "React" and "frontend development" are related — and to auto-generate personalized profile bios and smart connection request messages, making it easy for students to reach out without the awkwardness of cold messaging. Key features include a browsable skill directory with filters, a session scheduler with online/offline modes, a post-session rating system, and a personal dashboard tracking completed and pending swaps. Built with React, Node.js, and Firebase, SkillSwap turns every student into both a teacher and a learner — fostering a collaborative campus culture where knowledge flows freely between peers.
Inspiration
We noticed that students around us had skills others desperately wanted — yet no one knew who to ask. A CSE student struggling with design sat next to a design student who couldn't code. SkillSwap was born from that missed connection.
What it does
SkillSwap lets students list skills they can teach and want to learn, get AI-matched with compatible peers, and schedule sessions. Gemini API powers semantic matching, auto-generates profile bios, and drafts personalized connection messages.
How we built it
React + Tailwind for the frontend, Node.js + Express for the backend, Firebase for auth and real-time database, and Gemini API for all AI features. Deployed on Firebase Hosting with Cloud Functions handling Gemini calls server-side.
Challenges we ran into
Designing Gemini prompts that return consistent structured JSON for matching was tricky. We also had to handle free-form skill input — users type "reactjs", "React JS", "react" — and normalize them reliably before matching.
Accomplishments that we're proud of
Getting semantic skill matching to work in real time within a 6-hour window. The AI bio generator produces natural, student-sounding profiles instantly — something that genuinely surprised us during testing.
What we learned
Prompt engineering is a real skill. Small changes in how we framed the matching prompt changed result quality dramatically. We also learned to scope ruthlessly — a focused MVP beats a broken full product every time.
What's next for SkillSwap
Multi-college support, an in-app chat for session coordination, and a skill endorsement system. Long term — a resume-linked learning portfolio so students can showcase verified peer-taught skills to recruiters.
Log in or sign up for Devpost to join the conversation.