At a seed-stage AI biotech where design initially meant screens, I built the company’s first design function, and made its judgment work without me in rooms with no design precedent.

Domain Fluency

argued successfully in science and engineering terms

Domain Fluency

argued successfully in science and engineering terms

Power through Network Effects

understood and supported others’ priorities

Power through Network Effects

understood and supported others’ priorities

Evidence over Authority

led with actual user sentiment and behavior

Evidence over Authority

led with actual user sentiment and behavior

GenBio is building a virtual cell and companion lab interface, where scientists perturb cells computationally before committing to wet lab work. I joined as the first and only designer, on a globally distributed team of about twenty-five engineers and scientists. There was no design literacy, research function, or interaction framework; to most of the team, design meant UI.

While I was initially hired to make a founding concept real ASAP, the actual job was making design thinking a part of how engineers, scientists, and founders decided what to build and why. That came from two strategies: fluency in their domain and language, and bringing in real users as early as possible.

The path I designed for influence

Started with something concrete

I built a design system and shared language made to be inferable and extensible.

Engineers built from its logic; stakeholders had something real to react to.

→

↓

Used contrast and provocation

I designed all artifacts to elicit reactions that pulled the team out of blue-sky ambiguity.

Deliberately offered slightly wrong answers so the corrections surfaced the right ones.

→

↓

Brought users into the room

I designed and led a study with working scientists and implemented analytics early.

The team quickly used these insights to roadmap development.

→

↓

Made the thinking extensible

I created a decision framework from a 42-paper review, and development processes framed in engineering’s terms.

Founders used the framework in their meetings, and the Head of Engineering asked me to lead more process development.

In an environment with little structure and high ambiguity, I worked through relationships: engineers became my research partners, business development helped fund my studies, and I created a shared language to explain and argue design’s purpose in terms engineers and scientists could adopt. By the end, the whole team, founders included, was using my terms and framing to make decisions.

I build design functions with the intent that others can exercise good design judgment when I’m not in the room.

I build design functions with the intent that others can exercise good design judgment when I’m not in the room.

The throughline

In scientific AI, trust is key. Design can’t move the product forward without strong relationships, and users need multiple ways to verify outputs from the very beginning. I invest in both early, relying on evidence and research findings to provide transparency into my decision-making, and to make collaborators understand the user as a real person in the development process.