I design for the version of the problem that’s actually true.

Most product problems get solved twice: once for the version stakeholders describe, and once for the version the evidence actually supports. I try to spend as little time as possible on the first one. That’s most of what “opinionated” means when I say it about my own work.

I care about healthcare, AI, and SaaS specifically because they’re the domains where the constraints are real (regulatory, technical, organisational) and a design that ignores them doesn’t just look wrong, it doesn’t ship.

I use AI daily, for ideation, research, drafting, and debugging, not as a novelty but as how I actually work now. I expect the same of the products I design: if a tool doesn’t make someone faster at reaching the right answer, it’s not done yet.