I design with AI, build for AI, and ask the questions most people skip.
10+ years in B2B SaaS, biotech, and fintech — now building with AI and evaluating it before anyone trusts it with real decisions.
A local-first course on AI agents — because privacy architecture is a design decision, not an IT one.
On why your AI tool's privacy policy is a design decision
AI agents as a personal workflow layer, not a product feature
Full and Part time projects · AI product design · UX strategy
Design is a decision practice, not a delivery service.
The best design happens before anyone opens Figma. It happens in the conversations about what to build and why.
AI should make harm harder, not just work faster.
The goal is not eliminating risk. It is making harm harder and less normalized. That starts with knowing what a model can and cannot do before you hand it work that affects real people.
Strategy without ethics is just optimization.
Efficiency is not inherently good. The question is always: what does this make easier, and for whom? What does it make harder, and for whom?
Complexity is not the enemy. Confusion is.
I do not simplify by removing features. I organize complexity so the right information is available at the right moment for the right decision.
That's why I test before I trust. → AI Lab
Try to catch the magenta pixel.
Click when you're close. Some problems can't be solved by moving faster.
Click when you're close.