
Change one position.
Explore different shapes and chemical groups at one position. Study what each change does.
We are building AI models to predict how molecular changes affect binding and biological activity. That understanding guides the evolution of new drugs.
Strong binding alone does not make a useful drug. Our AI models are being developed to connect molecular edits to affinity, specificity, stability and biological response.

Explore different shapes and chemical groups at one position. Study what each change does.

Use shape, chemistry and physical forces to understand how a candidate fits and binds.

Aim to block a harmful interaction while preserving the proteins and functions the body needs.
Our approach combines AI models, molecular physics and computational directed evolution. Our drug candidates are varied, evaluated and selected over successive generations.
Every target calls for different selection pressures. We are training AI models on the relationship between a candidate’s shape and chemistry, its target binding site, and experimentally measured binding affinity. That understanding guides which changes to explore and which experiments come next.
Start with a target protein and the biological effect the drug needs to produce. Choose a binding site to explore.
Generation 01
Co founder, Illume Bio
sambhav@illumebio.net
Co founder, Illume Bio
kushal@illumebio.netIllume Bio is a drug discovery company building AI models for the scale of that challenge. Our ambition is to extend drug design to targets with limited binding data, using what we learn across programs to guide new candidates and the experiments that test them.