I build to learn. Along the way I picked up design, engineering, and product the way you pick up ingredients when you learn to cook. The goal stays the same: shorten the path from idea to something people use.
This is the approach we take at Womp and the reason I created KookieUI. It’s a working set of habits that helps us move with clarity and keep the product coherent.
AI moves the start line closer to the finish line. At Womp, it gives people a new on-ramp into 3D: generate an object in seconds, then place it, tweak it, and print it. Newcomers get momentum; experienced creators keep depth and control. Faster starts mean more cycles of taste, judgment, and intent—where the human work lives.
The person with the most context carries the work from idea to live. Graphics engineers ship the small frontend panels tied to rendering and experiments they run. Platform and backend engineers build thin UIs that sit on their services. I lead design, work in Figma, and make last-mile polish commits so frontend stays on core architecture and performance, and designers stay focused on problem-solving.
Clear ownership, quick reviews, and tight feedback compress the loop. Fewer handoffs. More progress.
We keep the interface consistent by putting UI rules in one place. I built KookieUI, a Radix UI fork tuned to how we build. Product code imports components like Button. Tokens, states, and interactions live inside the system. When spacing or motion changes, I update KookieUI and publish a new version; frontend bumps the package and the app adopts the standard automatically.
This separation keeps design decisions durable and lets engineers focus on logic and performance. At our stage I act as the design engineer—writing components, setting tokens, reviewing usage, and shipping updates through the package—so polish arrives with speed.
We release MVPs that cover the happy path and measure real pull. Interviews, dashboards, and usage patterns guide what grows and what stops. Reliability, performance, error handling, and smoke tests keep the bar steady. For AI-heavy areas, we preserve the quality of our generative features and help bot, maintain speed, and keep user intent clear.
A recent example: AI scene manipulation. Today it changes materials and moves objects with a small scope by design. We’re watching sessions and interviews to decide where to scale it next.
I run this loop end to end—timelines, product decisions, design work, and final touches in code through KookieUI—so the team can stay deep on the hard parts while the product keeps learning.
Start fast with AI. Turn context into speed through shared ownership. Keep pace and polish with a strong system. Ship small to find the pull. That’s how we build at Womp: fewer handoffs, tighter loops, and a product that stays clear as it grows.
