Output without understanding is just a faster way to build the wrong thing
henry tavarez · new york · product designer · research, systems, 0→1 · open to staff / principal / lead roles
I am the person teams bring in to figure out the right answers before building starts. Most of my work lives in complex, data-heavy places: public health surveillance, enterprise platforms, systems where a wrong guess costs something real. The UI is maybe twenty percent of the job. The rest is understanding people, and lately, teaching machines to carry that understanding without flattening it.
The work
Research, product thinking, and the design systems that hold them together. I have led design on public sector modernization where the users are epidemiologists and the edge cases are outbreaks, and on enterprise platforms where a thousand quiet decisions determine whether software feels built for somebody or for nobody in particular. I am comfortable in strategy and in the weeds, and I would rather cut my own work than ship padding.
The way I work now
I build with AI in the loop, and I am not conflicted about it. The bottleneck is not building anymore; it is knowing what to say before you hit enter. So the craft has shifted to where it always mattered: context going in, editing coming out, knowing what does not need to exist. This portfolio and Tessera were both made that way, with me directing Claude Code as the engineering agent. The thinking and craft stayed mine.
Right now
I am building Tessera, a thinking layer that turns product thinking into structured context an AI builder can actually use. It is in beta and free to try at tessera.app. I am also open to the right room: staff or principal product design, ux lead, somewhere the bar for craft is high and the questions are hard.