Designing AI that earns trust

AI-powered search, verification, and agentic workflows, designed to be explainable and controllable for intelligence analysts who can't afford a wrong answer.

Selected work

  1. Evolving Search interface showing the unified search bar

    Evolving Search for Intelligence

    Primer AI    From Boolean syntax to AI-interpreted intent for intelligence analysts.2024–25Case StudyAI AdoptionSearch UX
  2. RAG-V's verification pattern inside Primer's agentic search interface

    Making AI Verifiable

    Primer AI    Claim-level verification for every AI-generated statement.2024Case StudyVerificationScale
  3. A refined product dashboard mock lifted off an agent coding session building it

    Designing Against the Model

    Primer AI    Three questions about AI-native design, tested before trusted.2026–OngoingProcessAgentic WorkflowAI-Native
  4. UXAI framework cards

    Bridging ML Complexity and Trust

    UXAI    An explainable AI design framework co-founded at UC Berkeley.2020FrameworkExplainability
  5. Eightfold email modal with multiple follow-ups and scheduled send

    Enabling Efficient Outreach

    Eightfold AI    Multiple follow-ups and scheduled send for recruiter outreach.2019Case StudyEnterprise UXWorkflow Design