At Lemna We:
decode & encode
the
interactions
of life
Biological interactions are at the core of health and disease. We are building AI models that explain how molecules interact, from atomic-level interfaces to biological networks.
We use these models to identify high-confidence disease targets, optimal engagement strategies, and drug molecules that act safely and effectively.

Complex Biology,
Predictable Interactions.
Geometric Transformer at the Core
The geometric transformer is at the core of our models. It reasons from first principles using atomic structure, guided by biological context to uncover mechanisms of disease and drug action.
Most failures in drug discovery stem from an incomplete understanding of biological interactions. This leads to low-efficacy targets, ineffective drug mechanisms of action, or unforeseen side effects.
Our AI models biological interactions to generate explainable hypotheses that link atomic-level interactions with their functional biological context.
These hypotheses guide drug discovery teams in mapping disease pathways, nominating high-confidence targets, assessing druggability, and identifying molecules that engage targets safely and effectively.
Empowering a Full Cycle of Discovery
-
Pesto
Actively used by over 3000 researchers to predict interaction interfaces and molecular hotspots.
Use model (soon)Read paper -
Carbonara
Benchmarked by Novo Nordisk as one of two top protein sequence design models.
Use model (soon)Read paper
Unlocking a Golden Age of Biomolecules
We are a team of machine learning experts, biophysicists, bioengineers, and more based along the beautiful shore of lake Leman in Lausanne, Switzerland. 🇨🇭
We always welcome motivated and hardworking individuals looking for a role at the bleeding edge of biological engineering.
