You.com

• Improved chat experiences
• Shipped Docs V2
• Led design system initiatives
• Improved onboarding
• Cross-functional collaboration
• Codebase contributions

When I joined You.com, the company was expanding from AI search into enterprise agent experiences and developer-facing API products. I led design across agentic chat, DevEx, API docs, onboarding, and design systems—helping shape the developer and enterprise experiences that supported this transition. I also used Claude Code to accelerate UI refinement and vibe-prototyping, making it easier to explore and ship polished product interactions quickly.

I used Fern’s docs-as-code framework to customize the documentation site’s look and feel, partnering with Fern to resolve design architecture constraints along the way. I also worked closely with engineering to restructure the docs into a clearer, more seamless system of pages and tabs.

I explored and pitched an idea for an Invalidity Search Analysis Tool powered by a patent index. In the demo, users upload a patent and an AI workflow searches for prior art across patent databases and non-patent literature. I built a working prototype with Vite and FastAPI, then tested it against PTAB cases to evaluate retrieval quality. The prototype surfaced limitations in my initial setup, showing that a vector database over the USPTO public patent record would be the next step toward a reliable MVP.

When building web intelligence tools for agents, developers are hesitant to integrate APIs without confidence in the consistency and quality of search results. Even a few days of poor performance can create major issues for products that depend on accurate, reliable search experiences. To address this, we shipped a major upgrade to the Analytics experience, introducing clearer charts and operational metrics around request volume, error volume, success rate, and latency.

In collaboration with Engineering, Product, and Research, our five-person team dedicated a week to ideating, designing, and testing a new first-run experience to help enterprise users reach value faster and reduce early drop-off. We hypothesized that users were not fully aware of the company-approved data sources available through their organization’s platform administration. To address this, we designed a step-by-step onboarding flow that helped users configure their workspace around their goals and available connectors. We also tested “jumpstart” features that gave users immediate, relevant entry points into the product so they could engage without friction.