I grew up around people
who looked at the world differently.
Architects, artists, photographers. The kind of people who notice structure before surface.
I ended up in architecture school, which trained me to think before building — to understand load paths before picking materials, to ask what a space needs to do before deciding what it should look like.
That training didn't leave when I moved into product design. It just found a new medium.
A lintel is the element that spans an opening — a door, a window, a passage. It's what takes a building from sculpture to architecture. Without it, a structure has no way in, no light, no human interaction. It becomes useful only when something spans that gap and makes the opening possible.
That's the design I'm interested in. Not the impressive object, but the thing that creates real access for the people on the other side of it.
Expert users.
I've been designing software for 15+ years across industries where the stakes are high and the domain knowledge is deep.
Enterprise platforms, government systems, fintech, AI-native products. The common thread isn't the industry — it's the type of problem. Complex systems that serve expert users who don't have time for software that requires translation.
Select work shown publicly. Additional enterprise work under NDA.
legibility isn't useful.
The design problem isn't building AI products. It's building AI products people can rely on.
A lot of designers are treating AI as a production question — can it do the task? That's not wrong, but it's not the hard part.
The hard part is trust. And trust doesn't start when the AI delivers something — it starts the moment you ask someone to tell it something personal.
Koers, the cycling coach I built, opens with a conversation. Not a form. The AI is asking about your job, your family, your injury, the race you've been quietly thinking about. It's doing the work of intake — building a model of a person — but it has to feel like being heard, not processed. That's the first design problem. Get it wrong and nothing that comes after it matters.
The plan that comes out of that conversation explains itself. Every adjustment comes with reasoning in plain language. Not as a feature — because without it, the whole thing falls apart. You can't trust a coach whose thinking is invisible.
I'm looking for work at the intersection of AI capability and human expertise — products where the design problem is genuinely hard because the user is genuinely sophisticated.
Enterprise internal tools. AI-assisted workflows. Systems where the gap between what the technology can do and what the user can trust it to do is still wide open.
If that's the kind of problem you're working on, I want to hear about it.
Start a conversationSomething worth building?
The practice is small by design and runs on referral. If the work fits what you're building, get in touch. That's usually where it starts.