ABOUT ME
I design where people & AI actually meet.
Previously at ANNA Money, designing the internal tools a customer support team relies on, and increasingly, the AI systems working alongside them. My role isn't always about crafting pixel perfect designs, but streamlining operational workflows and processes building intuitive and automated systems that deliver the best outcome for our customers.
I sit between support and engineering, close enough to the front line to see where things break, and close enough to the code to help fix them.
I care about craft as much as outcomes, atomic components, clean tokens, interfaces that hold up at scale. But I'm just as drawn to the human side: watching where a support agent gets stuck, or figuring out how much of a problem an automation and where AI can own manual and repetitive tasks that create unnecessary work for support teams .
HOW I WORK
🔎
Discovery
Every project starts with understanding the actual problem, not the assumed one. I gather feedback and insights from the people closest to the issue, agents, users, engineers, data, and use that to define the real scope before anything gets designed.
🧭
Define
I turn discovery into a PRD: a clear, shared definition of the problem, the deliverables, and what success looks like. This becomes the reference point for everyone involved for the rest of the project.
💡
Ideate
I move fast into quick concepts, rapid prototyping or vibe-coding ideas into something tangible rather than staying in the abstract. The goal at this stage is momentum and options, not polish.
Critique
Ideas get shared early and often, with engineers, users, and stakeholders, so feedback shapes the direction before too much is locked in. I'd rather be wrong in a sketch than wrong in production.
Scope
With feedback in hand, I define the MVP: what's essential, what can wait, and how the project might be sliced into releases so value ships sooner rather than all at once.
Test
QA isn't an afterthought, I test against the original problem, not just the spec, to make sure what's shipping actually solves what we set out to solve.
Deploy
Ship it, often in slices, so value lands sooner and risk stays contained.
Iterate
It doesn't stop at deploy. I track metrics and gather feedback on what actually shipped, and feed that straight back into discovery for the next round. The process is a loop, not a line.