Variational AI - We discover novel drug-like small molecules

Redefining the unit economics of drug discovery for better patient outcomes

With an estimated 10⁶⁰ possible drug-like molecules, the chemical space is unimaginably large. And yet, most programs still start from the same limited chemical matter.

Traditional drug discovery begins from known compounds (screening hits, literature scaffolds, familiar chemical matter…) and then commits to long, costly optimization cycles, over-investing in the exploitation of limited starting points.

Variational AI believes in a different approach. Using generative AI, we design novel, synthesizable small molecules optimized across potency, selectivity, ADMET, and synthetic feasibility, giving R&D teams higher-quality leads to initiate their discovery programs.

Better starting points mean fewer costly iterations, faster timelines, and a higher probability of success, so more programs reach the patients who need them.

Our approach

From target product profile to synthesis-ready leads in weeks

Enki™ generates novel lead-like structures to meet selectivity, potency, toxicity, and pharmacokinetic program goals.

1. Define the preclinical target product profile (TPP)

Specify the On/Off-targets and physico-chemical properties of the molecules as input for Enki™ generative AI.

2. Enki™ generates compounds

Enki™ generates novel and diverse structures that meet the defined TPP.

3. Make your selection

Pick compounds you want to synthesize and test. Enki™ can also perform hyper-efficient lead optimization while constrained to a defined scaffold.

Built for early-stage drug discovery

Enki™ is purpose-built for the earliest stages of small molecule drug discovery — generating first- and best-in-class compounds for 873 pre-trained drug targets spanning kinases, GPCRs, proteases, oxidoreductases, hydrolases, ion channels, and other key target classes.

Generative AI, not screening

Enki™ generates novel molecular structures de novo — moving beyond fixed screening libraries to explore uncharted regions of chemical space.

Multiparameter optimization

Enki™ simultaneously optimizes across 50+ parameters — potency, selectivity, toxicity, ADMET, PK, synthetic accessibility, and more…

Learn more

The team behind Enki™

At Variational AI, we bring together machine learning researchers, chemists, and biotech professionals united by one mission: transforming early drug discovery with generative AI. Our cross-disciplinary team is always looking for new talents to push the boundaries of science and technology alongside us.

Meet The Team

See open positions