How I work
Start with a conversation.
Before I open Figma, I talk to the people who will use the thing. At Tendo, that meant sitting in on discharge conversations and watching CDI specialists identify comorbidities. At Avy, it meant interviewing knowledge workers and finding common patterns. I'm looking for where things break down, because that's where opportunity hides.
Find the real problem.
The first problem someone hands you is usually a symptom. Cardiology didn't need a better survey. They needed to stop making daily phone calls. Pay attention to what people say, and pay even more attention to what they do.
Make it rough. Test it early.
Get prototypes in front of real users fast. Every round is a chance to learn. Changes at this stage are inexpensive compared with production-ready code.
Stay in the room.
I work alongside PMs and engineers from the first conversation through launch. Some of the best design decisions I've made came from an engineer asking, "What if?" halfway through a build.
Ship, measure, go again.
A product isn't done when it ships. I look at what people actually do with it and use that to decide what comes next.
What about AI?
I was frustrated by how unpredictable AI-assisted development could be, so I built a small team of agents and a Kanban board for them to work from. Each agent has a role, every task is visible, and nothing ships until I've reviewed it. It lets me take an idea, turn it into a working prototype, and put it in front of real people through a process that is fast and predictable.