Work
None of them came with a clean brief or a controlled environment. The strategy work happened in triad meetings. The HITL breakthrough came because a backend team needed a builder and asked a designer instead of an engineer. The Deals System insight was three lines of text below a price. Most portfolios clean this up. This one doesn't.
A backend engineering team needed a front-end builder. They asked a designer. I'm the first non-engineer to push code to a Zoro production database. I'm documenting everything so others can follow.
Zoro's promo infrastructure was broken. We rebuilt it over 15 months, one phase at a time. The biggest win was a three-line text change: +27% revenue per user.
I wondered if I could prototype in Cursor instead of Figma. Turns out I could. Sixteen months of compounding later, my AI-generated code was running in production.
Aaron and I were the only designers prototyping in code. So we built a component library on our exact stack, wrapped it in skills, and made it so anyone can start prototyping in about 90 seconds. Nobody asked us to.
Zoro tried twice to build a design system and stalled twice. Attempt three is underway, built on the Storybook nobody asked for. Week one: all 772 production components, audited.
