Recent projects span consumer applications, enterprise workflows, AI-assisted products, data platforms, and production infrastructure. The work ranges from product definition and architecture through implementation, deployment, and operation.
Some projects are private, have been retired, or expose only a sign-in screen publicly. The descriptions and images below preserve enough context to show the work without exposing private systems or customer data.
An installable, camera-first team scavenger-hunt platform. Hosts create reusable hunts, invite players, form teams, release clues, review GPS-aware photo submissions, score results, and generate post-event recaps. The production system includes passwordless identity, private media, live event state, and AWS infrastructure.
An offline-first recipe PWA that stores normalized recipe data and imports recipes from pages, pasted text, or photographs. It combines structured-data extraction with vision and language models, while supporting search, favorites, variations, sharing, private media, and reliable source provenance.
A visibility platform for local businesses in the era of AI search. It discovers how businesses appear in chatbot recommendations, evaluates competitors and ranking signals, and produces actionable visibility reports. The system spans public marketing, commerce, operator tooling, and a multi-stage data and AI workflow.
A camera-based AI application designed to turn an everyday photo into an immediate, useful recommendation without requiring typed instructions. The platform supports multiple visual experiences, direct camera capture, AI image understanding, sharing, and a production cloud architecture.
A long-running commerce intelligence and operations platform built to identify product opportunities and improve marketplace performance. Proteus unified Amazon, Shopify, Walmart, advertising, inventory, review, search, and sales data; powered analysis and reporting; and automated recurring operational workflows at substantial scale.
A system for bringing overlooked public-domain books back to life as searchable, accessible digital editions. The broader platform combines scan reconstruction, constrained AI-assisted publishing tools, multi-domain static delivery, release manifests, editorial workflows, and recovery-aware AWS infrastructure.
Have a difficult product or platform problem that needs both strategic judgment and hands-on execution?
Talk about the project