loading project brian-ng
.jpg)
About Me
Practice
UX designer at Amazon. I prototype in production code rather than mocks, and I work with AI agents like a team of juniors — they do the grunt work, I direct and verify. The combination ships fast without sacrificing design-system fidelity.
Cognitive science and computing background from UCLA. Outside work: climbing, lifting, badminton. Strong opinions on AI integration and enterprise tooling, fewer strong opinions on bouldering grades.
What I Bring
Shipping under pressure
From zero context to shipped features in two months on Growth Conversations, four weeks for the approvals pattern for promotions. I move fast without sacrificing research depth.
Designing for scale
I build patterns that extend without redesign. The approvals pattern shipped for an initial use case, then extended to VP-level promotions, senior IC engineer promotions, and other Amazon products without structural changes.
AI-augmented practice
I use Kiro and AI tooling to draft research plans, build production-ready prototypes, and write technical handoff docs, compressing typical timelines while maintaining quality.
Experience
Amazon
UX Designer
Oct 2023 - Present
Amazon
UX Designer Intern
Summer 2022
How I Work
Ship a real prototype, not a perfect mock.
Production code in research sessions and engineering handoff. Engineering builds from it directly instead of translating from Figma.
AI is the IDE, not the architect.
AI agents do the grunt work — drafting, scaffolding, handoff docs. The judgment calls (synthesis, voice, design-system fidelity) stay with me.
Make the implicit explicit.
The structures experienced users already follow are usually the right ones. My job is to surface and formalize them so newer users get the same scaffolding.
Tight loops over long handoffs.
Reduce time between intent and shipped artifact. AI tooling and production-code prototypes both serve this end — and catch problems while they're cheap to fix.
Kiro IDE with five MCP servers wired up — Stencil for design-system fidelity, Figma for design context, Chrome DevTools for in-browser inspection. This is the setup behind the AI-augmented practice claim above: real prototype code, real agent sessions, real iteration.
Get in touch
Let's connect
Currently considering senior IC roles at companies building tools for HR, talent, or operations teams. Specifically interested in the parts of enterprise software where the leverage is high and the visibility is low.