Ship AI that works in production
Agents, reinforcement-learning data infrastructure, and on-device models. Designed, built, and shipped to real users, not left in a notebook.
I build products, end to end.
Applied AI and agentic systems, built on deep full-stack product work: on-device macOS apps, data infrastructure, and web platforms, shipped from seed to Series D.
Three ways I help teams and founders turn ideas into shipped, production systems.
Agents, reinforcement-learning data infrastructure, and on-device models. Designed, built, and shipped to real users, not left in a notebook.
Full-stack engineering across macOS apps, web platforms, backend services, and AWS, with the production discipline that keeps software dependable.
Architecture and modernization reviews, AI feasibility, and build-vs-buy guidance, so you invest in what is actually worth building.
The track record behind it, most recent first. One line per stop, the details live in conversation.
Core product engineering at a seed-stage automation startup, from client discovery through shipped features across the desktop app, integrations, and backend.
Built AI data infrastructure end to end: a reinforcement-learning platform for autonomous agents, dataset-generation tooling, and native capture apps for robotics training data.
Full-stack product and payments work: a transaction-driven donation platform, a migration from Stripe to Adyen, and a CRM rebuild that cut licensing costs.
Led the rebuild of a legacy account-processing system on Go microservices and AWS, drove containerization across multiple services, and built a large-scale consumer messaging system.
Performance and modernization work at an official NVIDIA partner: a new CDN architecture, legacy application migrations, and a guest checkout that lifted conversion.
Security product work at a company acquired by Cisco: role-based access control and incident investigation tooling across Go, PostgreSQL, and React.
Selected builds: a shipped macOS app, an AI learning tool, and a founder-era product.
Built and shipped a native macOS OCR utility in Swift, integrating an on-device model to extract text from any selected screen region, taken all the way from development through Apple’s review to the Mac App Store.
AI-powered learning application using a Llama 2-based model to generate questions from user documents, with a LangChain document-processing and retrieval pipeline and a WebSocket-connected React frontend for interactive generation.
Founded a location-based commerce platform in Go, Django, React, PostGIS, and Stripe, with a Python scraping pipeline ingesting product catalogs, geospatial search, a four-person team, and ~10 local businesses onboarded.
A toolkit sharpened across startups, scale-ups, and founder-era work.