Design System Engineering
Info
- B2B • Design System • Code connect
A scalable design system for a global data platform built with AI-native workflows.
Leveraged AI workflows to develop and deliver a scalable, unified design system for a global data platform.
Open full sizeProblem
Fragmented UI was stealing focus from real user problems.
The client’s data platform faced widespread UI inconsistencies across its ecosystem. Teams spent energy fixing interface discrepancies instead of solving user pain. The goal was to unify the visual language, improve accessibility, and reduce cognitive load.
This project is an independent revision in collaboration with AI agents, from an older project I’ve worked on.
Principles
Designing in the age of agents.
A practical framework for integrating generative AI into the design workflow — from insight generation to multimodal orchestration. AI is leverage in the system build, not the identity claim.
Process
Four phases. Each closes a step from insight to shippable system.
Discovery
We addressed UI inconsistencies to unify the platform, improving usability and accessibility to establish a shared visual language. This lessened the focus on UI alone, helping product teams better address user pain points. We also enhanced navigation and support access for users.
Tool stack
Research
By performing market research in collaboration with an AI agent, we identified established design systems for similar platforms. One strong reference was the Carbon Design System (IBM), especially for accessibility, consistency, and modularity.
Tool stack
Exploration
Reflecting on the challenge: how can we create a design system that’s simple, scalable, and understandable for designers and developers? We recognized developers’ needs and collaborated with a front-end AI agent to improve the system. We integrated tokens, CSS classes, and code snippets into the Figma file, enabling quick learning from visual harmony and coding perspectives.
Tool stack
Build
The system was crafted to expand and support numerous users. By applying our new design system principles, we enhanced readability for GenAI, utilized nested components, and optimized component properties for smoother variations.
Tool stack
Designs
Pattern docs and component specs carry the craft proof.
Open full size
Open full size
Open full size