Currently Design System Lead at PUMA, connecting design, brand, product, CRO & UXR, engineering, content, marketing, and AI, so a global brand can move fast and still feel like one.
Leading the design system of a global brand and scaling it to 25 countries across channels.
What it covers, how it is structured, and how it is governed.
Bringing web and app onto one shared visual language: the same tokens and components on every screen.
Consistent tokens, UI, and patterns across platforms.
Built a design system the team actually uses.
Founding an AI prototyping culture, and the operations that keep the system running as it grows.
Made prototyping with AI part of the team's workflow, turning ideas into working, on-brand prototypes in hours.
Grid/Products Card/Product Chip/FilterFigma Make prototypes become A/B tests, with Claude Code and Gemini writing the test code in the A/B testing platform, so decisions come from real user data.
Bringing AI prototyping into each role's daily work, so anyone with an idea can build it and test it.
Clear ownership, simple rules, and training and champions in each team, so AI-made work stays on-system. The work behind 90% adoption.
Shared libraries, prompts and Figma Make starting points built on the system, with results in speed, quality and tests shown to leadership.
Leading a dedicated design system team, collaborating with a diverse cross-functional team.
Collaborating with different teams to create the best possible UX, while managing expectations at every level.
Helping people adopt AI with confidence, so new tools strengthen their craft rather than replace it.
Prove the idea fast.
Test with real users.
Measure the impact.
Train teams and champions.
Part of everyday work.
Mentoring designers and engineers, and building a team that owns the system with me.
Hiring, onboarding and setting clear goals, so the team knows what good looks like.
Translating the design system and AI into business language, so leadership, brand and regional teams across 25 countries align on priorities and trade-offs.
Switching modes
AI-ready systems, scalable discovery and automated delivery.
↓How I made the system readable and usable by AI.
Clear naming, tokens stored as Figma variables, and documentation written for AI.
Lets AI tools like Claude Code read real components and variables directly from Figma.
My own MCP server that gives AI the rules, usage guidelines, and do's and don'ts.
Links Figma components to real code components, so AI uses the real code and doesn't invent new code.
export const Component =
z.object({
name: z.string(),
guidance: z.string(),
});## Button primary Always use the brand colour token. - Do: use brand tokens - Don't: campaign colours
{ "name": "ButtonPrimary",
"props": ["size", "icon"],
"variant": "primary",
"brandLocked": true }{ "button.primary.bg":
{ "$value": "{color.brand}" },
"button.primary.label":
{ "$value": "#FFFFFF" } }
Figma MCPCaptures design metadata from Figmasource: Figma
Design System MCPServes the code: components, tokens and rulessource: structured content
Claude
Cursor
Gemini
ChatGPT
Muse
CopilotA framework to scale A/B testing and UX research, turning ideas into live, on-brand tests in hours and scaling weekly testing to 6 times the usual pace.
Founding the AI prototyping culture. Turning ideas into working, on-brand prototypes in hours.
Prototypes move into live A/B tests & UX research in a fraction of the usual time, so ideas are proven with real users and data before they're fully built.
AI that lets product managers turn an idea into a working, on-brand prototype and a live test, so decisions come from real users and data before anything is fully built.
Every step uses the design system, so a test variant looks like the real product.
Research runs on realistic, on-brand prototypes, so users react to the real product, not a mockup.
AI for designers and developers, powered by one design system and a strong AI infrastructure, scaling both design and engineering delivery.
AI that scales design on the Figma canvas, combining different MCPs to raise design quality and scalability.
AI that scales design implementation in code with 80% accuracy, using Claude Code with Figma MCP, Design System MCP and Code Connect.
AI drafts with library components only, checks itself against the system, and the designer adds the craft.
Claude Code reads the real design, maps it to real components and follows the system's rules, so engineers review and iterate. With high accuracy, design-to-code implementation scales and more features ship faster.
The short version