Centralizing a design system at scale

TextCortex · Design Engineer

TextCortex is a mature product with a large dashboard. Products like that carry their history in the interface. Different areas get built at different times, by different people, under different deadlines. Each area is reasonable on its own. Together, they drift apart.

My work there had two threads: improving individual areas of the dashboard, and pulling the design system into one central place so those improvements would stick.

Improving the dashboard, area by area

I worked across different parts of the dashboard, tightening what was already there: clearer hierarchy, more consistent spacing, better states, fewer one-off patterns.

This kind of work rarely looks dramatic in isolation. No single change is a redesign. But interfaces are experienced as a whole, and a dashboard where every area behaves the same way feels calmer and more trustworthy than one where each page has its own rules.

One system instead of many

The deeper problem was structural. Design decisions lived in too many places: styles defined per-page, components duplicated with small variations, values that had no single source of truth.

So the bigger part of the job was globalizing the design system. Centralize the decisions, colors, type, spacing, and shared components, then make the product's screens draw from that one source instead of carrying their own copies.

The point of this is not tidiness. It's leverage. When a decision lives in one place, improving it once improves the whole product. When it lives in twenty places, every improvement is a project.

What it changed

Centralizing the system made consistency the default instead of a cleanup task. New screens started from shared parts, existing screens converged toward the same language, and the product got easier to scale, because building more of it no longer meant diverging more.