projects
Systems I have founded, built, or maintained, from assistive AR and AI-agent-driven hardware to mobile-network and open-source tooling.
AnySign
A bidirectional sign ↔ spoken language translation system for assistive AR glasses, built together with linguistic specialists, the Deaf community, and smart-glass vendors. It recognizes, translates, and produces sign language in real time, with all processing on the mobile device for immediate, low-latency communication. Now in field trials with Deaf communities in California.
Lets an AI coding agent take a hardware product from a plain-language idea to working firmware, a DRC-clean PCB, and a 3D-printable enclosure, and then on to a fab order. Seven agent skills cover firmware, script-generated KiCad PCBs, parametric CAD, industrial design, product-lifecycle contracts that keep the three domains in sync, and voice guidance for hands-on steps. Everything is generated by script and reviewed in the browser, so a beginner can ship a board without years of EDA and CAD experience.
The first open-source community tool for fine-grained mobile network analytics inside smartphones. It lets researchers collect and analyze cellular protocol messages directly on commodity phones, and has been adopted by researchers at more than 300 academic and industrial institutions.
AceMap
An academic search engine that maps the relationships among papers, authors, and topics from large-scale scholarly metadata. Its visualizations surface research trends and scholarly relationships, helping researchers navigate a field at a glance.
A small, readable JavaScript 3D library that makes graphics and WebGL approachable for learners, and the foundation for the course projects of more than 1,000 computer graphics students. Version 2 rewrites it on WebGL 2 with interactive examples, and it powers the demos in Graphics Foundations, my notes on the math, pipeline and algorithms of computer graphics.