Hack-Nation · Global winner
AMBR — AI for biology
Turning bacterial genomic data into a report that makes the model's evidence and uncertainty visible.
What we built
AMBR combines genomic annotations, machine-learning predictions, and calibrated uncertainty in an interactive report. The prototype can abstain when the available evidence is insufficient. It is a research tool, not a clinical diagnostic system.
My contribution
I worked on the product concept, frontend, research framing, demo video, and team coordination. We built AMBR together during the Hack-Nation Global AI Hackathon.
The team
Daniel Hofmann, Tom Burkart, Maximilian von Klinski and I brought different research and engineering backgrounds to the project. Our team won the challenge, the Harvard/MIT Hub, and the overall competition.