ABOUT ME

Think in systems — Build at scale

Hi, I’m Amy — a systems designer living at the intersection of design, humans, and AI.

Since I was a kid, I’ve learned and revised in systems. You would find a little version of Amy at a small table, synthesizing, summarizing in bullet points, drawing diagrams — until she eventually realized this is how she naturally learns, reasons, and makes decisions. Over the years, my knowledge has formed in systems and flows. I learn any software quickly because I see the flow, the logic, the mechanisms and patterns — not only the screens and the buttons. I am drawn to complex systems and how system logic connects to humans. I sketched princess characters at age seven, but my first design project was an ERP flow in Oracle — and two years later, a CRM system for an education center.

People build the houses and the buildings. I build the road — the infrastructure running underneath, the systems that everything else runs on, before a single screen ever exists.

Thinking in multiple layers with a high resolution range — I’ve carried this combination into every product I’ve deeply built. I take pride in creating beautiful products that people love, but I am more drawn to how context and cognitive load change users’ behaviour (“Don’t make me think”), the architecture layer — the wires behind the wall — and the governance layer: how changing one data point can affect the whole system.

My proudest projects live in the invisible layer: information hierarchy, navigation design, the design system, and workflow. I save months of engineering cycles, shorten both QA and design time, and hold the system out of drift.

In March 2026, I authored a workflow connecting Claude Code to Figma through MCP and the codebase, and rebuilt an entire design system from code. The process compressed months of work into weeks and removed nine years of friction between Figma and code. Code and Figma now mirror each other. This leads to something bigger: the ability to automate code and design both ways — I call this the bilateral pipeline, where AI moves fast in execution, and both developers and designers become the curators who hold the system out of drift.

Thinking in systems has become my main advantage in the age of AI. AI raised the cognitive ceiling: it demands structured documentation, good reasoning and logic, and the ability to communicate not only in context, but in systems. The way I approach AI is different. They are my collaborators — they think alongside me, stress-test my reasoning, and lend me their speed.

For me, AI is a new type of user. Reading these words, you may have realized it too: now that systems finally speak our language, we have all quietly started designing for AI. AI is becoming the main consumer of the design system — a user with its own constraints and cognitive load. For this new type of user, the question becomes: how do we build the ecosystem, the framework, the governance layer that supports them to work efficiently?

AI amplifies the input — so if the input is wrong, inconsistencies emerge everywhere, months later. One data point can structurally strengthen, or quietly break, everything downstream. A systems designer can identify the root of the problem: the font base changed to a percentage. The drift lies in a human structural decision — in the governance layer. It is not caused by AI, and it is entirely preventable.

AI moved the bar, but here is what I am sitting with: my judgement about what to build matters more now, not less. The risk is no longer building beautiful screens too slowly — it is shipping the wrong thing faster, and watching it break silently afterwards. I am here to keep the systems intact.

Most recently, I led design system architecture at Ratehub, including a Claude + Figma + Storybook MCP pipeline that compressed months of rebuild into weeks. Before that, I led platform navigation convergence across 100+ products at NielsenIQ, spanning 95 countries and serving Walmart, Loblaws, P&G, and 23,000 other B2B clients.

Outside of work, there’s a long tail

Some things made it — a math app for kid, the portfolio you’re reading, a few agent systems built just to see if they’d work. Some are queued and patiently waiting — an indoor plant that’s fine, food photos that stay in my camera roll, a social media account about AI and tech predictions with zero posts. Medium writing. The ideas arrive faster than the hours do. I consider this a good problem.