Microsoft Design System

Core Engineering is a Microsoft organization that develops products for internal customers. It covers hundreds of products—from lightweight experiences (intended for casual customers) to data-heavy business applications. The intention for many products built by the Core Engineering teams is to thoroughly test them by the “customer 0”—Microsoft employees—before launching products externally. With so many products, the design coherency went out of hand over the years resulting in some of them not behaving like a part of Microsoft family. It also led to inefficient product development teams building features addressing identical customer needs.

The Core Engineering Design System needed to move beyond atomic components and introduce advanced patterns and flows that improved efficiency and consistency across Microsoft’s internal products. Our objective was to systematize product development by building on Fluent, providing clear, reusable design patterns tailored to common workflows, and ensuring high accessibility standards. Additionally, we aimed to strengthen collaboration across design system teams and product groups to reduce redundancy and design debt.

Over time, CSEO products evolved independently, leading to inconsistent user experiences. Customers interacting with multiple Microsoft products had to constantly adjust their mental models when performing similar tasks. The Sankey chart visualization highlighted the fragmented use of Fluent components across teams, creating unnecessary friction. With hundreds of interconnected products and diverse perspectives across Microsoft, aligning teams around shared solutions required structured processes and collaboration frameworks.

We treated the Design System as a product, prioritizing efficiency and impact. Partnering with a researcher, I led customer discovery efforts to identify the most urgent needs of designers and product teams. Using the Jobs-to-Be-Done framework, we uncovered key focus areas: notifications (errors, push alerts, popups), learning, information organization, and grid structures. This informed our prioritization strategy and transparent processes, setting clear expectations for planning and execution.

Beyond establishing processes, I developed a framework that provided overarching guidance for design patterns. This allowed designers to focus on individual components while maintaining consistency at a system level. By collaborating with product teams, we ensured that solutions addressed top pain points and delivered meaningful impact.

As the accessibility design lead for the Design System, I worked to bridge the gap between legal requirements, technical implementation, and the design process. My goal was to make inclusive design a seamless practice rather than a compliance burden. I set accessibility standards, evangelized best practices, and created detailed interaction specs for all components. These guidelines were adopted by multiple teams, including the Fluent Design System team, and were recognized by Microsoft’s Central Accessibility team (CELA) for their contribution to improving accessibility across products.

I led Microsoft’s data visualization workstream, aligning 18 products and over 30 designers from across the company, including Teams, Azure, Power BI, and Dynamics. My role centered on establishing a unified approach to data presentation, synthesizing insights, and driving alignment across teams. By fostering trust and ownership within the design community, I ensured that new patterns were championed by the designers and organically integrated into their product work. The result was a successful design convergence that influenced Microsoft’s data visualization strategy and led to driving our point of view to code.

As the Lead Designer for MyHub web—an incubation project later launched as Microsoft Viva—I shaped the app architecture, interaction framework, and key user experiences. MyHub aimed to enhance employee productivity, connection, and satisfaction. I leveraged this project to validate and refine Design System patterns, incorporating real-world feedback and edge cases to strengthen the system’s adaptability.