# Kevin Chen (solo)

- **Event:** [Zero to Agent: Vercel x Deepmind Hackathon SF](https://cerebralvalley.ai/e/zero-to-agent-sf)
- **When:** Sat, Mar 21 at 9:00 AM – 10:00 PM (PDT)
- **Where:** Shack15, San Francisco, CA
- **Team:** [Kevin C](https://cerebralvalley.ai/u/k3vnc)
- **GitHub:** https://github.com/kvn8888/HackathonVercelGemini
- **Demo video:** https://youtu.be/jbMy5PUUWCI
- **Gallery:** https://cerebralvalley.ai/e/zero-to-agent-sf/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/zero-to-agent-sf/hackathon/gallery/55

CodeGym — an AI-powered coding practice platform that goes beyond LeetCode/DSA.

https://hackathon-vercel-gemini.vercel.app/

Stack: Next.js 16 + React 19, Supabase (Postgres), Google Gemini, Vercel Sandbox, Monaco Editor, BetterAuth

What it does:

Problem Generation — Gemini creates realistic, real-world coding problems (not DSA), complete with starter code, reference solution, test harness, and hints
Real Code Execution — User code runs in a Vercel Sandbox against actual test cases (with a Node.js eval fallback)
MCQ Marathon — Adaptive multiple-choice quizzes that target weak areas
Chat memory insert — Currently the functionality is limited to the user inserting context into the system for the memory feature (to manually add memories)
Skill Memory — A living AI-generated profile tracking domain scores (0-100), strong/weak skills, and interaction history. Every problem attempt, MCQ session, and chat updates the profile via Gemini
Core loop: Generate problem → solve it → execute against real tests → AI updates your skill profile → next problem adapts to your weaknesses.

---

Markdown version of https://cerebralvalley.ai/e/zero-to-agent-sf/hackathon/gallery/55. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt