Full-stack product engineer with 20+ years of turning rough ideas into shipped products, most recently as the sole engineer behind an AI-powered voice training platform. Equally at home architecting cloud infrastructure, building React/Next.js interfaces, and integrating LLMs, real-time speech and payments.
San Jose, CA · Open to in-office or hybrid in Silicon Valley, San Francisco, or remote.
I'm a full-stack product engineer with 20+ years of turning rough ideas into shipped products, most recently as the sole engineer behind an AI-powered voice training platform.
Career spans game QA at THQ, eight years as Director of Development at a digital agency for Shell, Maserati and Hawker Beechcraft, and founding eTono, a software studio. I'm equally at home architecting cloud infrastructure, building React/Next.js interfaces, and integrating LLMs, real-time speech and payments.
Comtac's firefighter training app puts trainees in command of a live incident without a human on the other end. Starting with a 360-degree structure walkthrough, the trainee calls in a size-up, directs arriving units, and works through the scenario verbally, with a dispatcher and incoming crews respond in real audio, keeping the incident moving in real time. At the end of each session, a self-evaluation and scored review let the trainee measure their performance against the call.
I built the web app end to end. The pipeline captures mic audio and transcribes it with ElevenLabs, sends the transcript to the OpenAI API to generate contextually appropriate dispatcher and unit responses, then synthesizes them back to speech through ElevenLabs. Incident video and the 360-degree walkthroughs are hosted and streamed through Mux; voice recordings are stored on Amazon S3; session data, self-evaluations and scores are saved for post-session review. The platform integrates with Stripe for recurring subscription payments.
Comtac's firefighter training app puts trainees in command of a live incident without a human on the other end. Starting with a 360-degree structure walkthrough, the trainee calls in a size-up, directs arriving units, and works through the scenario verbally, with a dispatcher and incoming crews respond in real audio, keeping the incident moving in real time. At the end of each session, a self-evaluation and scored review let the trainee measure their performance against the call.
I built the web app end to end. The pipeline captures mic audio and transcribes it with ElevenLabs, sends the transcript to the OpenAI API to generate contextually appropriate dispatcher and unit responses, then synthesizes them back to speech through ElevenLabs. Incident video and the 360-degree walkthroughs are hosted and streamed through Mux; voice recordings are stored on Amazon S3; session data, self-evaluations and scores are saved for post-session review. The platform integrates with Stripe for recurring subscription payments.
Honolulu, HI

