Case Study · Indeed · Employer Product
Embedding within employer product teams to build tooling, automation, implementation workflows, and operational systems that drove adoption from 45% to 94% across 500+ teams.
The problem
The centralized system solved foundational architecture.
But employer product teams still faced:
The issue wasn’t component quality.
The issue was: implementation infrastructure.
Product teams needed systems that worked inside real development workflows.
Embedding into product teams
Instead of operating purely from centralized systems, I embedded directly into employer product teams to understand where adoption was failing.
This meant working inside:
This helped identify friction such as:
This role became the bridge between centralized architecture and product execution.
Employer-specific product packages
Employer teams repeatedly rebuilt the same product-level patterns.
To solve this, I helped create employer-specific packages that extended the centralized system with reusable product infrastructure.
These included:
This reduced duplication while keeping teams aligned with central standards.
Automated quality systems
Manual design reviews couldn’t scale.
We introduced automated quality systems directly into engineering workflows.
Design system lint rules prevented misuse.
Chromatic and visual diff workflows helped catch regressions early.
Automated accessibility checks reduced QA burden.
Custom audits identified fragmentation and non-system usage patterns.
Pipeline tooling ensured safer updates.
The goal: move quality enforcement earlier into development workflows.
Design system observability
One of the biggest gaps was visibility.
Teams often assumed adoption was healthy when implementation quality was poor.
We built reporting systems to track:
This created much better operational visibility.
AI implementation tools
As AI coding tools became more common, I helped define how AI agents could operate safely within design system constraints.
This included:
The goal was allowing teams to move faster without introducing new inconsistency through AI-generated code.
Feedback loop into centralized systems
I wasn’t just consuming the design system. I helped evolve it.
I acted as the liaison between product teams and the centralized design systems organization by:
Community & evangelism
Tooling alone doesn’t solve adoption.
I also helped build:
This helped scale behavioral adoption alongside technical adoption.
Outcomes
Other projects
Indeed
Multi-platform architecture: semantic tokens, scalable theming, and governance that scaled adoption across hundreds of teams.
Read case studyIndeed
Evolving the design system for conversational and generative AI: interaction primitives and workflow architecture so hiring products iterate without fragmenting the platform.
Read case studyIndeed
AI sourcing with explainable fit, implemented on top of shared employer UI and workflow standards.
Read case study