Amazon

Amazon Personalize and Amazon Forecast brought advanced machine-learning capabilities to AWS customers without requiring them to build the underlying infrastructure themselves. My work translated complex models and workflows into products that customers could understand and use.

I saw an opportunity to give Amazon’s warehouse software a coherent design foundation. I built the case for the system, secured funding, and led the work from its initial direction into a shared foundation for Amazon Fulfillment Technologies.

The system had to support a demanding product ecosystem: different device classes, languages, accessibility needs, screen sizes, and physical working environments.

Warehouse software could appear on a 5-inch handheld device or a wall-mounted display as large as 105 inches. The system gave teams a consistent foundation across that range without treating every screen as the same interface.

Accessibility and localization were built in from the beginning. We also introduced dark mode to reduce eye strain and respect the conditions in which employees used the software.

In 2020, we designed left-to-right and right-to-left behavior into the system itself. Components, layouts, and interaction patterns could move between writing directions without requiring each product team to redesign its software.

At the time, that level of bidirectional support was unusual at the design-system level. It made localization a property of the foundation rather than a downstream retrofit.

The result was a shared design foundation built for Amazon’s fulfillment network: accessible, localizable, adaptable across devices and display sizes, and considerate of the employees using it every day.