I built and maintained a design system used across multiple products - then spent as much energy getting teams to actually use it as I did building it.
The Challenge
Design systems fail organizationally, not technically. The real work is making the right choice also the easy choice for every consuming team.
The Impact
Over 30 components shipped and adopted across multiple products, transitioned from sole ownership to team-contributed model without losing architectural consistency.
Automated tools tell you whether ARIA is valid. They don't tell you whether a keyboard user can finish the flow - that gap is where we put the architecture.
The Challenge
Complex dashboards, multi-step forms, and confidential data flows for users who may rely on screen readers or keyboard navigation - treated as an architecture constraint, not a post-build audit.
The Impact
Shipped WCAG 2.1 AA compliant, mobile-first portal serving as the primary interface for people filing disability claims.
A pseudo-element on a single list item costs nothing. On thousands of layers during a drag, with live updates coming in, that's a different problem.
The Challenge
Thousands of layers with drag-and-drop reordering, scrolling, and real-time collaboration on hardware less powerful than today. Users expected native-app fluidity from a web-based design tool.
The Impact
Frame time improved 30-40% across both JS and CSS/DOM optimizations. Users noticed without being prompted - the fix addressed a known pain point.