I've been meaning to start writing for a while. Working on AI products day to day, I keep running into questions that don't have clean answers in papers or docs — they live in the messy space between research and production. This blog is where I'll try to write some of those down.
What I'll write about
Mostly things I'm actually working on:
- RAG in production — what breaks at scale, evaluation strategies, and the painful gap between a working notebook and a system real users rely on.
- Voice AI — VAD, streaming STT/TTS over WebSocket, and the latency budget that makes or breaks a conversation.
- Medical LLMs — guardrails, hallucination handling, and why "looks confident" is the most dangerous failure mode in healthcare.
- The boring stuff that matters — observability, deployment, evals, and team workflow.
Why a static blog?
I wanted something I could write in plain markdown, push to a tiny Python script, and be done. No CMS, no database, no auth — just files on disk and an nginx config. The whole site is under 500 lines of hand-written HTML and CSS, and that's the way I like it.
The best tool is the one that gets out of the way.
If you found this and want to chat about AI engineering, voice systems, or anything in between — send me an email.
More soon.