Location: San Francisco, CA, USA

Remote: Yes (remote or hybrid)

Willing to relocate: Yes (SF, NYC, Austin, other US hubs)

Technologies: Python, TypeScript, Java, C++, FastAPI, NestJS, Next.js, React, Node.js, PostgreSQL, SQLite, Docker, AWS, PyTorch, Hugging Face, TRL, LangChain, LangGraph, LLM APIs (OpenAI/Anthropic), RAG, fine-tuning (GRPO/DPO/LoRA), multi-agent systems, observability, REST APIs, Git, CI/CD

Résumé/CV: https://darshannere.com/resume | https://github.com/darshannere

Email: darshan.nere2@gmail.com

Most AI agents in production are black boxes. When one fails, nobody can tell you why. I build the tooling that fixes that.

Tracea (tracea.dev): open-source observability for AI agents. Traces every LLM call by patching the HTTP transport layer, so there are no SDK wrappers, no framework lock-in, no code changes. Install and you see everything.

ObservAgent: local-first observability dashboard for Claude Code sessions. 1000+ active users, open source.

Before this: shipped a civic complaint platform to 50k+ users for a municipal government.

Looking for SWE engineer or AI engineer roles at startups building real production AI. High ownership, real users, hard problems.