The people building Living Models.
Three people based in Lyon — a computational genomicist, an ML engineer, and a plant geneticist. We built this because the genomic data that could predict trait performance from sequence already existed; what didn't exist was a model architecture that could use it.
Computational genomics background from the Université Claude Bernard Lyon 1 and a post-doc at INRAE Auvergne. Spent five years working on genomic selection pipelines for wheat before concluding the label-scarcity problem was the fundamental bottleneck — and that foundation models were the right approach.
ML engineer with a bioinformatics background from ETH Zürich. Designed the k-mer tokenization architecture and the multi-species joint training pipeline. Previously built high-throughput genomic analysis infrastructure at a European clinical genomics company, processing WGS data at population cohort scale.
Plant geneticist with 15 years in experimental breeding at INRAE and subsequently at a major European seed company. Leads experimental validation partnerships and is the bridge between our model outputs and what breeders can actually use in a selection cycle.
Scientific advisory board
Three researchers from outside the company who review our validation methodology, challenge our evaluation choices, and keep us from overstating what the model can do.