Inside My Process

Moving an idea from “what if” to “here, try this.”

I’m not a traditional engineer. I turn observation into working prototypes by directing AI-assisted development — which is what lets a single person take an ambiguous idea to a tangible, testable artifact in days, not quarters.

  1. 01

    Observe

    Two decades on the floor of Apple stores is two decades of watching real people meet new technology. The raw material is human behavior — what people reach for, resist, misunderstand, and quietly depend on — not a list of features.

  2. 02

    Frame

    Reduce a big idea to its smallest buildable piece. PetTag before the whole of Attune. A narrow thing you can actually ship beats a grand thing you can only describe.

  3. 03

    Prototype

    Build it through AI-assisted development — Claude Code, Codex, GitHub Copilot, ChatGPT — across web and visionOS. I direct the architecture, interaction, and product decisions and assemble working software fast.

  4. 04

    Test

    Put it on a real device or deploy it, then react to how it actually feels. RoomBridge lived on Vision Pro; Between is live on the web. Feel beats spec.

  5. 05

    Iterate

    Keep the loop short. The prototype isn’t just a demo of the idea — it’s how the idea gets discovered and refined.

Tools

AI-assisted development: Claude Code · Codex · GitHub Copilot · ChatGPT.
Web: Next.js · TypeScript · React · Supabase · Vercel · GitHub.
Apple platforms: iOS · macOS · visionOS. Design: Photoshop · Illustrator.

The honest version

I direct and assemble working software through AI-assisted development rather than hand-writing production code. For an incubation team that uses prototypes to discover ideas — not just demonstrate them — that speed from observation to working artifact is the point.