I'm Ramona. I design interfaces and build them myself, so nothing gets lost between the mockup and the browser. Motion, states, accessibility and all.
A design system built from one token file: two token layers, thirteen components, and a build that fails if anything drifts. The travel-ops dashboard is built entirely from it, documented in Storybook in light and dark.
Live earthquakes from the last month, explored by asking questions in plain English, like "deep quakes near Fiji." The AI only turns your question into a filter. Every dot on the 3D globe and every number comes from the USGS feed, so nothing is made up.
Motion you tune by eye, not by number. Springs are named for how they feel, every token races its siblings in a live playground, and what looks right exports as JSON, TypeScript or CSS.
Can a quiz feel like flipping through photo prints? AI writes ten questions live, each paired with a photo.
Does teaching sell tea better than selling does? A brand site with a working steep timer.
Loading, empty, and error get the same code review as the happy path, each one designed and built before QA ever sees a build.
Every color, radius, and easing curve lives in a variable. The design survives refactors because it was never hardcoded in the first place.
Motion curves picked per component in code, never the library default. Hover the stage to feel one.
Agents scaffold components, explore variants, and surface the states I'd forget. Every line still gets read before it merges.
I started in front-end engineering, turning Figma files into production code and finding everything the mockups didn't say, like loading states, awkward breakpoints and focus rings. So I started designing those parts myself, then whole flows, then the systems behind them. Now I own an interface from layout and states through the shipped component, and I can explain every decision in it.
