Memorang is the AI infrastructure stack for education, trusted by Cambridge and ranked #1 in the Vercel AI Accelerator. We redesigned it end-to-end across five products, cutting educator review time by 47% on a platform now serving 200M+ assessments.

YC S24-backed, raised a $4M oversubscribed seed led by Race Capital with our pitch deck, with customers preventing 7,000+ data leaks and reducing AI costs by 70%. We've been the founding design partner for 20 months across three product pivots, shipping the design system, every product surface, the Framer marketing site, and the deck that closed the round.

Superorder is backed by Foundation Capital, Michael Seibel and Y Combinator, Kyle Vogt and Daniel Kan of Cruise, and a roster of operators who collectively represent some of the most demanding product taste in the industry, a $10M Series A built on the thesis that restaurants deserve the same operational infrastructure as any other scaling business. We came in as an embedded design partner replacing a high-cost US studio, and the job was specific: take a platform entering hyper-growth with a non-scalable UI architecture and build the component-based design system it needed to handle the next phase without the interface becoming the bottleneck.

Previously led deep learning research at the world's most advanced AI labs, now building a Python SDK that turns runtime exceptions into auto-generated pull requests. We led end-to-end design for a founder with no design background and shipped a parallel Claude Code prototype of the full platform that accelerated his build speed 10x.