We teach LLMs biology.

Biology runs on context.

Most AI in biomedicine excels at proteins and molecules. But disease emerges across hierarchies: gene circuits within cells and communicating cells across tissues. cells integrate cytokines, chemokines, surface contacts, spatial cues—and talk back. Traditional datasets and cell-line assays flatten this complexity. Our insight: to understand and ultimately program biology, models must learn the combinatorial language of cellular context.

The cellular operating layer.

• Read

  Learn how cells interpret multi-signal microenvironments.

• Write

  Generate programmatic interventions and virtual screens in context.

• Generalize

  Multimodal pretraining across omics, imaging, and literature.

• Verify

  Design validations that translate to ex vivo human systems.

cell110 trains foundation models aligned to biology's hierarchies and grounded in human-relevant data. The models forecast context-dependent responses and support the programmatic design of interventions—moving beyond isolated targets to tissue-level understanding.

Context-native foundation models for living systems.

• Multimodal pretraining

  Integrate omics, imaging, knowledge, and literature.

• Context graphs

  Represent interacting cells and signals over space-time.

• Programmatic outputs

  Virtual screens and generative design in tissue context.

• Uncertainty & OOD gates

  Calibrated predictions for translational decisions.

From mechanism to intervention.

Results, not correlations.

We prioritize prospective validation in ex vivo human systems. Our models have identified context-dependent mechanisms that were subsequently confirmed in the lab, demonstrating the ability to forecast effects that cell-line assays miss. Ongoing collaborations expand external validation across indications.

Collaborating with leaders in AI and biomedicine.

We work with select research groups, pharma R&D organizations, and AI labs to advance contextual modeling and validation.

Selected collaborators under NDA; references available upon request. Academic & AI research; top-10 pharma; leading institutes.

Name origin.

cell110 bridges biology and computation. The "110" references Rule 110, a well-known elementary cellular automaton capable of universal computation: from a few simple rules, it generates rich, emergent behavior. We use cell to anchor our mission in living systems. Together, cell110 expresses our philosophy: simple biological rules, executed in context, give rise to the complexity of life—and by teaching models this language, we can predict and program cellular behavior.

Build the operating system for biology.

We're assembling a small, exceptional team across foundation models, representation learning, computational biology, and translational science. If you've shipped state-of-the-art models or engineered hard biological systems—and want your work to matter—let's talk.

Open roles

• Senior Research Scientist, Foundation Models (multimodal)
• Research Engineer, Spatiotemporal Modeling
• Computational Biologist, Perturbations & Context
• ML Infra Engineer (training & eval)
• Partnerships Lead (R&D collaborations)

Partner with us.

For pharma collaborations, AI data partnerships, or research briefings, share a few details. We'll respond promptly.