Designing Intelligence

The Manifesto

I started as a musician. Mastery was never the end point, it was the condition that made real expression possible. Twenty years of product design later, across design agencies and in-house teams at Mercedes-Benz R&D, Amazon, Apple, and Google, that idea still runs through everything I make.

At Google, I built internal tooling for the first Gemini launch and earned a named contribution in the landmark Gemini 1.0 technical paper. I now work on AI infrastructure and tooling.

The gap between what AI can do and what people will trust it to do is the most important design problem of our time. It won't close through better models alone. It closes through designers who understand both the technical and human sides, and have the craft to build the bridge between them.

That is the work. That is what I do.

Four principles for
AI product design

01

Probabilistic UX

Certainty is a UX convention. AI breaks it.

Design for confidence, not certainty. Communicate honestly, handle ambiguity, and keep users oriented when the AI is not sure.

02

Human–AI Collaboration

AI should make humans more powerful, not passive.

Automation that strips out judgment creates liability. Design for amplification: the right information, legible recommendations, and humans kept where they add value.

03

Invisible Guardrails

Safety that interrupts gets bypassed.

Build constraints into the interaction itself. Guide people toward better outcomes without breaking flow, so safety becomes a design quality instead of a checkbox.

04

Latency as Interface

The wait is a design canvas.

AI thinks before it responds. Design the wait to feel intentional, transparent, and useful: what is happening, why it takes time, and what comes next.

Thinking out loud.

Prototypes and essays on the craft of building AI people trust.

Experiment · 2026

Transcendental Particles

An interactive field of particles shaped by orbit, wind, and flow.

Open →

Prototype · 2026

Trading the Whistle

Building PITCHFLO around anticipation, momentum, and shared ritual.

Read →

Article · 2026

The Screen That Knows What You Need Before You Ask

Designing proactive AI for shared screens without taking action away from people.

Read →
See all

Latency as interface, applied

Start with intent.

What brings you here?

Choose a starting point and I’ll draft a note for you.

A starting draft

Recruiting for a role. Here’s a note to make your own.

Open in email ↗