MVP Development · Medical AI

A team of data scientists and medical researchers came to us with a working AI model, a clear vision — and no product. My job was to turn their research into something a clinician could actually use: a structured MVP where user experience, business requirements, and technical constraints all pull in the same direction.

The riskiest thing about research-driven products is that everyone sees a different product in their head. So before any design, I ran a six-hour workshop with the full group — the client's team and the developers together. We identified the key users, their goals and pains, built three personas with core scenarios, and mapped the journey-to-be for each user type. Just as importantly, we left the room with a shared list of features ranked by user value and technical feasibility — which meant the hard scoping arguments happened on day one, not in sprint three.

From there, we worked in weekly cycles with the client's team: user flows, concept sketches, and a prioritisation map that turned "everything is important" into an actual MVP scope. The discipline here wasn't generating ideas — it was cutting them. Every feature had to earn its place against the question: does the first version need this to prove the concept?

I designed medium-fidelity wireframes and built a clickable prototype — not a pitch deck, but something the team could put in front of real users and stakeholders and get honest reactions to. The prototype became the shared reference point for business, medical, and engineering conversations.