AI infrastructure for drug metabolism and pharmacokinetics
DMPK.AI is a technology company building machine-learning models for ADME, DMPK, and pharmacokinetic prediction — designed for pharmaceutical scientists and the developers who support them.
Mission
Make rigorous, calibrated DMPK prediction accessible at every stage of drug development — from the first virtual screen to translational modeling.
Scientific grounding
Models are built with practicing DMPK scientists in mind: familiar endpoints, honest uncertainty, and explainable structure-property relationships.
Engineering discipline
Every model is versioned, benchmarked on scaffold-split data, and served through infrastructure built for production use.
What we build
We develop and maintain a growing catalog of models covering absorption, distribution, metabolism, excretion, CYP450 interactions, clearance, and pharmacokinetic and pharmacodynamic behavior. Models are trained on curated public, licensed, and proprietary datasets, validated with scaffold-aware splits, and delivered through a consistent API and Python SDK.