Audit the current product UI and existing design artifacts; identify the highest-leverage opportunities to standardize, consolidate, and reduce inconsistency.
Define the information architecture of the system itself — naming, taxonomy, component hierarchy, and documentation structure — so that both humans and AI agents can find, reason about, and correctly apply the right primitives.
Create a concrete step-by-step plan by partnering with design, engineering, and leadership, to correctly prioritize pieces of this effort, balancing long-term scalability with near-term product needs.
Define and ship a scalable token strategy (color, typography, spacing, elevation, motion) with a pipeline that keeps Figma variables, code tokens, and Storybook in sync.
Partner with Front-End engineers to translate the system into reusable, well-architected components with clear APIs, states, and accessibility built in.
Own the component lifecycle end-to-end: versioning, release notes, deprecations, and handoffs.
Build a lightweight, builder-friendly contribution process — designers, engineers, and PMs should be able to propose, review, and add to the system without a heavy governance tax.
Create documentation that people (and agents) actually use: guidelines, do/don't examples, usage rationale, and decision records, structured so they're machine-readable as well as human-readable.
Drive pragmatic adoption: start where it matters most, ship iteratively, and build credibility through visible product impact.
Define how AI agents help audit, tag, and maintain the system's structure — and where human review gates belong (naming decisions, deprecations, new pattern introductions).
Shape the pipeline that takes a Figma frame or a PM's prompt through Cursor / Claude Code into Storybook-validated components, including the review checkpoints that keep quality, accessibility, and brand integrity intact.
Make the system legible to agents. Ensure components, tokens, and patterns are documented, named, and structured so that when a designer or PM prompts an agent to "build the post-intent screen," the agent reaches for the right primitives and produces something shippable — not something that needs to be rebuilt.
Define what success looks like and track meaningful signals: adoption, UI consistency, design/dev cycle time, accessibility improvements, and the quality of AI-generated output against the system.
Partner with Research and Data where useful to connect system changes to product outcomes (quality, user confidence, conversion efficiency).
Requirements
Has owned a design system end-to-end at a product company, including the organizational work of getting it adopted — not just the design artifact.
Deep fluency with design-to-code integration: tokens pipelines, component APIs, versioning, and keeping design and code in sync in practice (not just in theory).
Track record of driving adoption through pragmatic, builder-friendly processes rather than heavyweight governance.
Strong working knowledge of modern front-end concepts (tokens, components, states, responsive behavior) and comfort collaborating closely with engineers on tradeoffs.
Genuine curiosity and hands-on experience with AI-assisted design and development workflows — you've used tools like Cursor, Claude Code, or equivalent, and have a point of view on where humans need to stay in the loop.
Sets direction for peers and partner teams. Comfortable making opinionated calls on ambiguous tradeoffs (governance vs. speed, system purity vs. product reality, human review vs. agent autonomy) and defending them with evidence.
Ability to thrive in ambiguity and build structure where it doesn't yet exist — pragmatic, opinionated, iterative.
Clear communicator who can align diverse stakeholders and influence without formal authority.