Learn the interaction.
Design the molecule.
Frontier AI, grounded in biophysics and trained on the target itself. Metric learns directly from target-specific experiments to design molecules for the exact behavior the biology requires.
The design space is broad. The qualified space is small.
Therapeutic programs rarely need binding alone. They need the right combination of affinity, selectivity, conditional behavior, species reactivity, and developability. Each added requirement narrows the molecular solution space.
A selection campaign should produce more than hits.
Hit-selection workflows use campaign data to rank the sequences they observe. Metric uses the campaign to learn an interaction model and guide design beyond the observed hits.
Metric designs beyond the observed hits and toward the required molecular profile.
Built across physics, biology, and machine learning.
Metric was founded by physicists, biologists, and machine learning researchers with backgrounds at Columbia and Caltech. Our platform grew from peer-reviewed work by members of the founding team and is now being applied across multiple pharmaceutical companies and programs.
Bring us a hard interaction profile.
We work with teams on molecular design challenges where selectivity, conditional binding, species reactivity, and other interaction constraints matter.