I build for the messy, high-stakes stuff.

I’m Shahrose. For about a decade I’ve built software for operationally complex, heavily regulated corners of the world that don’t forgive sloppiness, mostly healthcare. I started in hospital operations, spent a few years building health-tech platforms at Verily, then four-plus years as the first product hire at Sprinter Health, working across the whole product, from the patient-facing apps to the care programs we sold to national health plans, as the company grew from a seed-stage startup to a $55M Series B.

I work best when someone hands me an ambiguous, high-stakes problem and gets out of the way.

I’d rather ship in days and learn from real people than spend three months perfecting a spec. I’m not an engineer by training, but I lean technical: I like scoping directly against the codebase, and I have strong opinions about where AI belongs in a product and where it quietly makes things worse.

Almost everything I do now runs through AI. I built my own agentic system, Jarvis, to run my work and my life, and I use it every day. I’ve put LLMs into production, and I care most about the unglamorous parts: evaluation, consistency, and knowing where the model breaks. What I want to do next is point all of that at healthcare, using AI to cut waste and actually improve outcomes for the people who need it.

Outside of it, I’m usually training for some race I’m underprepared for, hunting for a cheap overland rig to fix up and take somewhere with no cell service, or tinkering with a homelab that does considerably more than it strictly needs to. I build small tools for problems only I have. I take photos. And I’m slowly trying to get better at writing in public.

If any of that overlaps with what you’re into, or you just want to talk shop about AI and healthcare, I’m easy to reach.