How I Work
AI tools are a core part of how I work, not a side experiment. Below is a snapshot of how I use them day-to-day — at work and on my own.
At work
Full-time product designer
Built a shared GitHub environment where designers and PMs can build and test prototypes independently, without needing a developer. Removes the bottleneck between design ideas and working interfaces.
Use the Figma MCP to pass designs directly between Figma and Claude Code, keeping design and implementation in sync without manual handoff.
Created a Claude-powered skill that lets engineers validate their implementation against design standards on their own, without waiting for a designer to review. Shifts quality left in the process.
Use Claude Cowork to pull and synthesize user research data and product analytics, turning raw data into design direction faster than manual analysis.
Use Claude Design to create product demo videos for feature launches, used in emails and social media as part of the go-to-market flow.
Built a prototype in Figma Make showcased by the CEO at Coaches Congress.
Led design of the PushPress AI Assistant from v1 through a full rebuild. Defined the interaction standards, guardrails, and proactive AI framework (Trigger, Inform, Recommend, Automate) that the engineering team builds from. Shipped v2 to a mid-90s success rate.
Built out the design system's AI component library: Message, Conversation, Task, Suggestions, Prompt Input, Sources, and Chain of Thought — with usage rules and accessibility guidance adopted across product surfaces.
Personal projects
Built my portfolio site using Claude Code. Designed and deployed independently, no developer.
Wellness app for women in perimenopause, built solo using Claude Code. Shows how I move from concept to deployed product with AI tools, without a developer.
AI tools I use regularly
I take on one fractional client at a time.