Built to learn difficult molecular interactions directly.
Metric was founded by physicists, biologists, and machine learning researchers with backgrounds at Columbia and Caltech. Together, the team brings decades of experience in biophysics, experimental biology, and large-scale machine learning systems.
Single-cycle molecular discovery.
We are building toward lead-independent, single-cycle molecular discovery. Our goal is to start with a target and molecular profile, learn the interaction through one information-rich pooled campaign, and reach a validated candidate panel without a supplied lead.
One cycle consists of one pooled campaign followed by one locked validation panel.
The design bottleneck is moving.
Frontier molecular AI is becoming increasingly capable at proposing plausible binders. High-throughput experiments can now measure target-specific interactions across enormous sequence spaces. Metric connects those advances through biophysical modeling and target-specific training.
From published principle to proprietary platform.
Peer-reviewed work by members of the founding team established the core inference principle behind Metric. Since the 2022 publication, we have built a proprietary target-trained molecular design platform that extends the work across campaign design, assay-aware modeling, molecular search, and validation.
We are now applying the platform with biopharmaceutical innovators across immunology and oncology, including a leader in TCR–pMHC therapeutics.