Designing a System That Scales

Building a unified design foundation across products, teams, and code.

Role

Senior Product Designer - Design System Lead

Team

Design & Engineering - cross-product, 8 apps

Scope

Org-wide infrastructure - tokens, components & governance

a design system cover

My Role & Impact

I built and owned this design system as a product - from the initial Tailwind audit and token architecture, through the Figma–Storybook parity model, to the governance process that kept it from fragmenting as it scaled to 8 apps. I designed the contribution workflow, wrote the documentation standards, and mentored the design team away from screen-drawing toward pattern-based thinking.

  • Owned: Full system - tokens, components, governance model & team adoption

  • Drove: Figma–Storybook parity strategy & the "Three-Way" contribution governance

  • Outcome: 3× faster design-to-dev · 100% visual consistency across 8 apps · accessibility standard from day one

100%

Reduced inconsistencies

3x

Increased design-to-dev speed

8

Scaled system across products

The Challenge: Systemic Fragmentation

As the company evolved into a multi-product ecosystem, the UI layer remained fragmented. Design decisions were made directly in development using a raw Tailwind library without a centralized design strategy.

This created three critical risks:

  • Accessibility Liability: Inconsistent contrast and focus states were a major barrier for our primary users, who often navigate with visual disabilities or assistive technology.

  • Brand Dilution: A disjointed user experience across products, undermining trust in a sensitive service area.

  • Engineering Debt: Developers were "reinventing" the same components across different products, slowing down delivery and increasing maintenance costs.

  • Lack of Parity: There was no shared "language" between Figma and code, leading to friction during handoffs and QA.

Inconsistent visual language across products

Inconsistent visual language across products

Strategy: Bridge, Don't Replace

My strategy was to build a system that felt like a natural extension of the developer workflow, not a hurdle.

  • The Tailwind Audit: Instead of replacing the engineering foundation, I audited the Tailwind library and extracted only the essential components. This ensured immediate "dev-friendliness" and high adoption rates.

  • Foundations-First: I prioritized Variables (Tokens) over components. By defining the "DNA" (colors, spacing, typography) first, we ensured that every future component would be accessible by default.

  • Alignment over Aesthetics: I focused on aligning cross-functional teams on a Governance Model to ensure the system would grow through process, not improvisation.

System Work: Infrastructure over Interface

I treated the design system as a product, focusing on technical architecture that mirrors the codebase.

  • Accessibility as a Foundation:

    • Typography: Selected Inter for its high legibility and x-height.

    • Color Variables: Created a semantic naming system. Every token was validated against WCAG 2.1 AA/AAA standards to support veteran users.

    • Focus States: Engineered high-visibility focus rings specifically for keyboard-only navigation.

  • Design-Code Parity: I structured the Figma library to mirror the Storybook architecture. If I named a variable brand-primary-hover, the developer saw the same token in their code.

  • Variable Strategy: Implemented Figma Variables for spacing, radii, and colors, allowing us to test "Themes" (like High Contrast Mode) instantly across all 8 apps.

a figma components hierarchy
a storybook components hierarchy

Figma and Storybook

Figma and Storybook

Delivery & Collaboration: Driving Adoption

A system is only successful if people use it. I focused on building a "Contribution Culture."

  • Tailored Communication: I presented the system differently to each audience:

    • To Engineers: Focused on tokens, props, and reducing custom CSS.

    • To Managers: Focused on speed-to-market and cost reduction.

  • Governance Loop: Established a dedicated Teams channel for component requests. Every update was a partnership between a dedicated designer and an engineer to ensure the system remained lean.

  • Education: Mentored the design team to stop "drawing screens" and start "building with patterns," shifting the team’s maturity from execution to orchestration.

Management & The Contribution Loop

As the system grew from a "form-kit" to an enterprise-wide asset supporting 8 applications, the biggest risk was fragmentation—teams either creating "rogue" components or over-complicating the existing ones. I established a formal governance model.

1. The "Three-Way" Check

To prevent the design system from becoming a "junkyard" of one-off components, every request via our Teams channel was filtered through three questions:

  1. Is it Universal? Does more than one application (out of the 8) need this?

  2. Is it a Pattern or a Component? Can this be solved by combining existing components (e.g., a Card + a Button)?

  3. Is it Accessible? Does the proposed change maintain our strict contrast and focus-state standards for veterans?

2. The Request Lifecycle

I designed a clear workflow to ensure no request disappeared into a "black hole," which maintained high developer trust:

  • Discovery: A developer or designer submits a request in the Teams channel.

  • The "Sync" Meeting: A 15-minute technical review between the Lead Designer (System Owner) and Lead Developer to discuss API props and Figma constraints.

  • Drafting: The component is built in a "Beta" Figma file and a "Experimental" Storybook branch.

  • Validation: Final QA check for accessibility (keyboard tab-order and screen reader labels).

  • Promotion: The component is merged into the Main Library and documented.

3. Documentation & Parity

We enforced a "Twin-Track" documentation rule. A component was not considered "Done" until:

  • In Figma: It utilized Variables for spacing/color and included "Usage Guidelines" (When to use vs. When not to use).

  • In Storybook: It mirrored the Figma naming convention exactly (e.g., Input/Veteran-High-Contrast). This eliminated the "translation layer" between design and code.

a figma component table
a figma component page header
a design system preview

Selected screenshots from Figma

Selected screenshots from Figma

Key Outcomes

  • Engineering: 3× faster design-to-dev handoff - components became "drag-and-drop" rather than "custom-built," eliminating the translation layer between Figma and Storybook.

  • Engineering: Scaled to 8 applications with 100% visual and functional consistency - the system absorbed new product needs without fragmentation, validating the governance model.

  • User: 100% of new features accessible for veterans from day one - WCAG 2.1 AA/AAA compliance and high-visibility focus states baked into the token layer, not retrofitted.

  • Org Impact: Became the organisation's single source of truth - Figma and Storybook are now the reference point for PMs, QA, and stakeholders, not just designers and engineers.

  • Org Impact: Shifted team maturity from execution to orchestration - designers moved from drawing screens to building with patterns, measurably improving the consistency and speed of the whole design org.

Other projects

Thinking about your next design hire?

Let's have a quick conversation. I'm currently open to Senior Product Designer and Product Design Manager roles. If you think there might be a fit, a 15-minute call is a good place to start.

Thinking about your next design hire?

Let's have a quick conversation. I'm currently open to Senior Product Designer and Product Design Manager roles. If you think there might be a fit, a 15-minute call is a good place to start.

Thinking about your next design hire?

Let's have a quick conversation. I'm currently open to Senior Product Designer and Product Design Manager roles. If you think there might be a fit, a 15-minute call is a good place to start.

Copyright 2026 by Wojtek Boniaszczuk

Copyright 2026 by Wojtek Boniaszczuk

Copyright 2026 by Wojtek Boniaszczuk