Open Application

Outpost Bio pairs high-throughput functional assays of human-derived microbial communities with causal machine learning to build predictive models of microbiome-compound interactions. Unlike black-box approaches, our Lab-in-the-Loop process tightly couples experimental design and AI development, so the models inform the science and the science informs the models. We generate large-scale perturbation datasets using stool-derived communities (SDCs) and multi-omic measurements, then train frontier models that predict metabolism, toxicity, and response at the community level. Headquartered in Boston and London, backed by Seedcamp, Merantix, Defined, and Openseed VC.