Nine years designing AI products, healthcare tools, and spatial interfaces. I work where clarity isn't optional — it's the entire product.
Agentic AI, hospital scheduling, radioligand therapy training — each project put me in a system where users needed control over something complex, under real pressure.
Customers on the redesigned platform now report up to 40% traffic lift and 25–50% sales gains. I rebuilt a black-box, four-agent system into one users could steer — adding human-in-the-loop gates, dual-panel chat, and direct editing so AI output became a draft users shape, not a finished product they accept.
iF Design Award 2026 winner. A healthcare scheduling system built so the 2 AM shift call doesn't need a phone tree. I owned the manager core and the system architecture that connected desktop scheduling and nurse mobile — two platforms, three designers, shipped in seven months.
Launched at ASTRO 2025 and showcased at Novartis Meet the Management in London — recognized by the Novartis CEO. I led the design of the VR experience within Novartis's broader Radioligand Therapy program — 4 clinical spaces, 148+ storyboard frames. One interaction language built in Phase 1 and reused through Phase 3, so HCPs could practice radioactive protocols safely before touching a patient.
Nine years of shipping products that people depend on. Not portfolio exercises — real systems, real consequences, real users under real pressure.
The systems I work on don't get second chances. A wrong screen in an operating room, a confusing step in an AI workflow — these aren't minor annoyances. They're the moments design exists to prevent.
HIPAA-aware healthcare, regulated pharma, WCAG AA — I treat regulatory constraints as a design material, not a checklist someone hands you at the end.
Real systems run on old hardware and browsers nobody chose. I design for what's actually there, not what I wish was.
The less training a product needs, the better the design. I build interfaces that teach you as you go — quietly, without getting in the way.
When four AI agents make a decision together, someone should be able to understand why. I design the window into the machine, not just the output.
I prototype in code, synthesize with Claude, and reach for image generation when a brief calls for visual range I don't have to produce by hand.
I prototype interactions in code before opening Figma. This portfolio, every working demo I take to stakeholders, every quick proof of concept — Claude Code is faster than wireframing and produces something engineers can actually read.
Research synthesis, copy iteration, edge-case stress tests, design reasoning. The design partner I run my thinking by — for every clinical protocol, every interview transcript, every persona I'm trying to sharpen.
Standard Figma for delivery; Figma Make compresses hours of layout exploration into minutes when I'm scoping a direction. Useful for presenting three concepts to a client where I'd previously have shown one.
Health/Wellness UX, Desktop Software
Invited juror
Student work
Student work