Today I lead design for Microsoft Specialized Cloud AI with a team of ten designers: AI for customers who run Azure on their own infrastructure, at the edge, and even fully disconnected. I own Infra AI, and the platform experience across all three AI lines.
It started with the infrastructure itself. On Sovereign Private Cloud, I mapped how single products grew into an ecosystem across cloud, edge and on-prem, and designed Azure Local Disconnected Operation end to end, for places that can never connect to the cloud.
When agents arrived, I was the only designer on an eight-week, CVP-sponsored squad building multi-agent orchestration, designing how people see what agents are doing and stay in control. That work moved into the Microsoft Copilot product line and earned two Microsoft patents.
AI changed how I design, too. I started our team’s AI-native design practice, from a research agent to prototypes shipped as pull requests, and took it to 558+ colleagues: the studio, plus the PMs, devs and other stakeholders we work with.
I write about AI-native design judgment on Substack, and speak about it at Microsoft and at product conferences.
Before Microsoft, I led a design squad at IBM on Cloud Pak for Data, designing identity, permissions and diagnostics for data and AI platforms. It is where I learned that the hardest part of a complex system is making it legible.