Analysis time
87% reduction in time spent per plate
User satisfaction
Unanimous positive feedback from researchers across all testing rounds
Adoption
Live and in active use at Institut Pasteur de Montevideo
Process
I worked directly with the researchers at Pasteur throughout the process — sitting with them, watching how they worked, understanding what slowed them down and what they actually needed from a tool like this. The requests were specific: faster batch uploads, smarter ways to flag when the AI missed a plaque, multi-select for similar values, customizable formulas for results calculation, and collaboration across different user roles. Every one of those made it into the product. Prototypes were tested and validated with the team, iterated until the interface felt as natural as the workflow it was replacing.
Execution highlights
The full app covers study and project repositories, plaque review and approval flows, AI-assisted recognition with manual override capability, batch file uploads, real-time result calculation with customizable formulas, CSV export, and multi-user collaboration with role-based access. Designed for desktop, built to handle the complexity of a scientific workflow without ever feeling like it.
Role & Team
Sr. UI Designer at Arionkoder — UX research, full app design (desktop), prototyping, and user testing. Working alongside a Product Manager and an ML Engineer, in direct collaboration with scientists from Institut Pasteur de Montevideo.
Timeline
April 2025 · February 2026
Stack
Figma





