Full-stack developer with a focus on AI systems — agents, MCP servers, and RAG pipelines. I build things end-to-end: typed frontends, Python/Node backends, and the AI layer that connects them to real tools and data.
I care about the decisions behind the code, not just making it run.
Languages
AI / Agents
Backend
Frontend
Infra
|
Full-stack AI agent system with modular architecture. GitHub Agent for repo analysis, Email Agent via Resend, Anime Agent for profile-based recommendations. FastAPI backend orchestrating Groq LLM calls, Next.js + TailwindCSS frontend, MongoDB persistence, OAuth via GitHub and Google. Deployed on Vercel. |
MCP server that bridges LLMs to external tools: Pinecone (vector memory), Gmail, Google Cloud, Excel, and IDE connectors. Async tool handling via FastMCP. Designed to extend any local or cloud-based agent with enterprise-grade integrations without modifying the agent itself. |
|
Local RAG agent — 100% offline, no API keys. Semantic chunking via |
Production-oriented user management system. Node/Express REST API with JWT auth, Zod input validation, Prisma ORM over MongoDB Atlas, and pagination. React 19 + Vite frontend. API deployed on Render, frontend on Vercel. |
