Selected work
The products and the decisions that shaped them.
Healthcare, defense, and government. Each write-up covers what changed, why it changed, and what happened next.
01 · Product Definition
Field research · Workflow maps · Product hypotheses
Find the problem before the requirements harden.
The useful research happens where the job happens: clinicians between visits, analysts on shift, operators in the middle of an incident. What people actually do is usually better evidence than what they say they need.
Definition establishes the bet.
02 · Product Building
Functional prototypes · Design systems · Production code
Build enough of the real thing to learn.
You can’t learn how an AI agent should behave from a static screen. I build the working version: real components, real states, real data, on the same design system production uses. The prototype is part of the research, not the presentation.
Building makes the bet testable.
03 · Product Validation
Success criteria · Shadow runs · Go or no-go
Decide what survives.
Before features are shipped, I define what would make us stop: a number, a threshold, and a date to check it against. Then the idea meets real conditions rather than a curated demo, so the answer arrives while it is still cheap to act on.
The work is the decision, not the launch.