About~/about

I move fast from problem to shipped solution: I scope an open-ended problem, build the first working version end to end, then keep iterating based on real usage. I have production experience deploying self-hosted open-weight LLMs (GPT-OSS, Qwen, Gemma) and shipping LLM-powered agents that replace multi-person manual workflows. A full-stack background (Next.js, Node.js, PostgreSQL) means I ship AI features into real systems, not notebooks. In the open, I'm building Raggie, an open-source RAG toolkit, and I serve on the board of Linux Evi.