Somesh Kumar
AI Engineer | LLM, RAG & AI Agents | Full-Stack
New to Contra
Somesh is ready for their next project!
On the web side I build full-stack products end to end. kumarsomesh.com is my own: a Next.js frontend, an admin panel, a FastAPI backend, and an Android app, all built and deployed by me. I also shipped Dutch, a Splitwise-style expense splitter, and Prismock, a self-hosted...
A lot of AI features don't need the cloud. I build them to run on-device, so the data never leaves the user's machine. Audient records a meeting and summarizes it locally, Whisper transcription plus a local LLM, zero data egress. imlistening does real-time offline dictation on...
Most chatbots give wrong answers because of retrieval, not the model: what gets pulled in before it ever answers. So I built a RAG chatbot that answers straight from your own documents, with sources you can check. Qdrant vector store, FastAPI + LangChain backend, clean chat UI....
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.
Automated Code Testing system (client: Thomson Reuters). A Dockerized Python and Next.js tool that ingests a GitHub repo, auto-generates test cases, and uses LLMs to localize and debug failures, surfacing errors and edge cases to cut manual code-review cycles. Selected in week...
