Product engineer in Los Angeles building applied AI products, evidence-rich interactive systems, and creative software.
I take products from an unclear problem to a working, testable system—researching the domain, shaping the interaction, building across the stack, and checking where real-world constraints push back.
I’m especially interested in AI that supports judgment without replacing it: systems with traceable evidence, explicit human approval, and honest boundaries between a prototype, a validated build, and a released product.
- Job Search OS — a local-first system for discovering opportunities, evaluating fit against candidate evidence, producing truthful application materials, and keeping consequential actions behind approval gates.
- Inner Avatar — a reflective AI platform exploring consent-based memory, source transparency, governed knowledge, and reasoning visibility.
- ProgressionLab — an AI-assisted harmony and songwriting workspace with playable progressions, practical voicings, visualization, MIDI/PDF export, and sharing. Try the live app.
- Wave Factory Essentials — offline audio plug-ins built with C++ and iPlug2, including Threefold Palm multiband dynamics and Valley Spirit pitch-shifting delay. Available as pre-release AU, VST3, and CLAP tester builds.
- Flight Tracker AI — a read-only aviation-intelligence portfolio project combining aircraft, weather, trajectories, and explainable attention signals on one navigable map.
- EMF Visualizer — an interactive 3D laboratory for exploring electromagnetic emitters, interference patterns, and contested RF environments.
- Product design and engineering stay in the same loop.
- AI output is a proposal, not an authority.
- Evidence, provenance, consent, and approval are product features.
- A polished demo and a dependable product are different milestones.
TypeScript, React, Next.js, Node.js, PostgreSQL, OpenAI APIs, Rust, Axum, C++, iPlug2, Python, and the testing and observability work that makes these systems dependable.
Most of the work here is product exploration in public: some projects are deployed, some are actively being built, and some are deliberately bounded prototypes.