Workbench
2026- Product Design
- Design Engineering
An AI-powered tool for designing warehouses, with 3D simulation to test and optimize layouts.
- 2Fortune 500 design partners
- NVIDIAOmniverse partnership
An AI-powered tool for designing warehouses, with 3D simulation to test and optimize layouts.
GXO’s shared AI platform for managing warehouse sites, clients and orders.
A dashboard for raising capital and a checkout for investors to buy shares.
A warehouse network for same-day delivery, with each item tracked by RFID.
I work on problems that show up messy.
Most of what I build is software people use all day to get their work done: moving inventory, raising money or running an operation. Often, people have built habits around what’s broken, and the team has strong opinions about a process no one can fully explain. I work with the people doing the job to untangle it, then build software they don’t have to fight.
I lead that work with product and engineering, from setting a direction through to the details of what ships. Along the way, I build design systems that help the team keep the product consistent as it grows.
I started as a developer and still take ideas from prototypes to real production code. I’ve also shipped my own web, mobile and desktop apps and VS Code extensions.
When I’m not building, I’m usually rock climbing.
01Framing & collaboration
I work with users, product and engineering to understand the problem. We weigh user needs, business goals and technical limits, then agree on priorities.
02Prototyping
I use working prototypes to test the parts we’re unsure about. Seeing an idea in use helps us spot gaps, compare options and decide what to build.
03Systems & craft
I build reusable patterns and work through edge cases, then stay close to implementation. I review what’s built and refine it in code so the design holds up in use.
How does John use AI in his work?
John uses AI throughout design and development:
Exploring ideas
He uses it to explore layouts, draft copy and generate audio, images and video, then develops the ideas that fit the project.
Building with the design system
He builds components and full screens with AI, using the project’s existing patterns and design system to guide the work.
Making room for refinement
Tasks like sample data and first drafts of edge-case copy take less time, leaving more room to test and refine the product.
He sets the direction, reviews the output and decides what ships.
He uses Orca, Codex and Claude Code to build, and Claude in the browser to work through research and open questions.
He also uses tools like ElevenLabs and Higgsfield for audio, image and video generation.
Reusable skills and custom agents help with repeat tasks and code review. He keeps the code and project notes in GitHub so changes are easy to follow.
With AI, he’s built and shipped design systems, screens, features and full user flows, as well as advanced canvas-based tools.
AI also helps check for unused code, duplicate components and styles that should use shared design tokens. He reviews the changes and refines the details before they ship.
He starts with a clear brief: the problem, constraints, references and patterns already in the product. That gives the work a direction before any code is written.
From there, it’s a design review: check the layout, copy and details, keep what works and rework what doesn’t. The first output is a starting point.