Featured Project · 01
LaCuriosity
A fully automated editorial platform publishing daily long-form articles across five thematic streams — Science, Psychology, Technology, Mathematics, and Ecology — each driven by a specialized Claude instance with its own persona, research mandate, and visual identity.
The pipeline handles everything without human intervention: demand-signal research via Tavily and Reddit, multi-agent drafting and fact-checking, Risograph-style image generation with Gemini, voiceover via ElevenLabs, video assembly via Creatomate, and distribution through Ghost CMS, email, and social. One daily trigger. Fifty-five nodes. Zero human touchpoints.
Featured Project · 02
Notebound
An iOS app that turns a trip's photos, notes, and voice into a written story and a mapped, day-by-day itinerary. The opposite mode of agentic AI to LaCuriosity: human-in-the-loop, multimodal intelligence embedded inside a polished consumer product.
Capture is voice-first and offline-first. A multi-model Claude pipeline runs server-side — Haiku classifies intent and tags notes, Opus looks at the photos, writes the narrative, and emits a structured itinerary with coordinates.
AI Agentic
Engineer
I build in two modes of agentic AI. One is autonomous orchestration — a trigger fires and a chain of models research, draft, fact-check, generate, and publish with no hand-holding (LaCuriosity).
The other is intelligence embedded inside a shipped product — voice capture routed through intent classification into multimodal generation, working offline-first on a phone in someone's pocket (Notebound). The craft is the same: choosing the right model for each job, chaining them across APIs, and wiring it so the whole thing holds together under real conditions.
YouTube · Google
Mountain View, CA
Building systems that scale to billions at YouTube — and shipping AI products from scratch on the side.
Experience
Education
Skills
Recognition