Somesh Kumar
AI Engineer | LLM, RAG & AI Agents | Full-Stack
New to Contra
Somesh is ready for their next project!
Dutch, a Splitwise-style expense splitter I built in Next.js. Track who owes what across a group and settle up, with clean multi-user expense logic and real state handling. A focused, shipped full-stack product. Stack: Next.js, TypeScript, React.
0
20
My portfolio platform, built and deployed end to end. A Next.js frontend with its own admin panel, a FastAPI backend, and a companion Android app, all on Vercel and Cloudflare with real subdomain routing and SSL. It's the clearest proof I ship complete multi-part products, not just localhost demos. Live at kumarsomesh.com. Stack: Next.js, FastAPI, Python, Tailwind, Vercel, Cloudflare.
0
22
Audient: a privacy-first meeting summarizer that runs fully on-device. It records a meeting and transcribes, diarizes, and summarizes it locally, so the audio and the summary never leave your machine. Built with ElectronJS and FastAPI, pairing Whisper transcription and Pyannote diarization with local SLMs for zero data egress. Stack: ElectronJS, FastAPI, Python, Whisper, on-device SLMs.
0
38
RAG Chatbot that answers straight from your own documents, with sources you can check. Built on a Qdrant vector store with FastAPI and LangChain. The trick to accuracy is retrieval, not the model: get what's pulled in right and it stops making things up. Stack: Python, FastAPI, LangChain, Qdrant, Next.js.
0
35