2016-2019
Yale, 2016–2019 — computer science and mainstage musicals.
I studied computer science at Yale and produced mainstage musicals there from 2016 to 2019, engineering aerial elements along the way. One of them, Reverie, put robotic arm holograms onstage; ideas like it are now emerging in Broadway immersive shows. A 2017 idea for personalized spatial sound, placing museumgoers in the center of the action, was later featured in the 2024 re-launch of the Yale Peabody Museum.
2019-2020
Beijing, 2019–2020 — teaching STEM at AndKids.
After Yale I moved to Beijing and taught STEM at AndKids, building a maker community of students and parents around 3D printing, robotics, and coding. The cohort broke records for science and reading improvement. Teaching those students still fills me with hope that all learning is possible for every person, and there's joy in being part of others' journeys.
2020-2024
Humana, 2020–2024 — safe rails for AI products.
From 2020 to 2024 I was a senior product manager in Humana's incubation lab, advising the CIO and Chief Strategy Officer on AI deployment risks and incubating the software rails for repeatable, safe launches of AI products. Embedded with ML scientists and clinical pharmacists, I turned medication-refill risk research into a predictive capability social workers use to prioritize high-risk patients sooner. I also mentored two bootcamp students from their foodservice jobs into full-time product engineering roles in 32 weeks.
2024-2025
Healthyr, 2024–2025 — clinical care by app, SMS, and voice.
From 2024 to 2025 I was the principal product manager at Healthyr, where our flagship app delivered gamified clinical interventions to 70,000 patients and drove the daily behavior behind clinically significant reductions in metabolic syndrome. We ran live experiments every week; winning protocols went from prototype to production in under a week while protecting patient safety and system integrity (HITRUST). We pivoted to SMS and a real-time voice LLM, and rural usage grew to ten times the industry average.
2025
OTseek, 2025 — zero to one with bond traders.
In November 2025 I co-founded OTseek in Pear's PearX W26 batch, and we went from customer discovery to production with institutional bond traders. We navigated bank and hedge fund compliance and negotiated production NASDAQ connectivity. I built our agent-native engineering discipline, embracing Codex and Claude Code without producing slop, through property-based testing, formal methods, pair programming, and weekly retros with real human writing.
2026
OTseek, 2026 — a ChatGPT app in the first public wave.
In 2026 we went zero to one a second time, with US local government customers, and OpenAI approved our native ChatGPT app in its first public wave. In June 2026 we wound down amicably with our investors.
2026
July 2026 — between trains, shipping open source.
Up Next > NYRR Midnight Run > NYE 2026
It's July 2026 and I'm between things, building. I maintain a platform that is the infrastructure and guts of everything I build, and a CLI that makes maintenance, multi-agent collaboration, and building more enjoyable for me. I'm building HealthMCP to connect personal health data to AI with greater privacy and control, and incubating Runsetta to share the motivation and connectedness I feel through running. Making with AGI feels better to me when it's like sitting down for a bit to continue chiseling on marble, rather than never leaving my seat so I can be ready to pull the slot machine's lever.