One AI platform. Three critical R&D decisions.
Screen
Virtual compound screening
Predict cellular responses to compounds your lab has never seen. Screen millions of counterfactuals computationally, run physical assays only on the highest-conviction hits.
Safety
Cross-species transfer
Translate biological signals across species. Map animal and in-vitro evidence directly to predicted human toxicity, and kill unsafe programs before they enter the clinic.
Translate
Clinical outcomes
Connect preclinical interventions to patient evidence. Responder signatures, resistance mechanisms and biomarkers to guide patient selection and trial design.
Trained on your data. Deployed in your environment.
Pre-trained on the world’s largest repository of single-cell, spatial and bulk RNA data, then fine-tuned, securely, on your proprietary assays and historical outcomes.
Deployed fully on-prem or in your VPC. Your data never leaves your environment, and is never sent to third-party APIs like OpenAI or Anthropic.
Demonstrated predictive power.
The underlying architecture of CellType has already identified novel, physically validated biology.
To prove its predictions hold in living tissue, the model was tasked with evaluating 4,266 oncology compounds. It prioritized 23 candidates, leading to the physical validation of 3 novel hits in ex-vivo human tumor tissue.
Built by the inventors of Cell2Sentence at Yale, in collaboration with Google DeepMind. Published at ICML.
Published science.
The models behind CellType come out of peer-reviewed work on single-cell biology, language models and causal inference.
The team.
David van Dijk, PhD
Co-founder & CEO · Yale
Yale Professor of Medicine and Computer Science. 11,000+ citations. Published in Nature, Cell, NeurIPS and ICML. Invented Cell2Sentence.
Ivan Vrkic
Co-founder · EPFL, Yale
Co-developed Cell2Sentence at Yale, published at ICML and NeurIPS. Led large-scale foundation model training in industry. Built software to control CERN’s Large Hadron Collider.
Lucas Carey, PhD
Chief Strategy & Partnerships Officer · Pompeu Fabra, Peking University
Led R&D at Pando Bioscience and an ML team at Ginkgo Bioworks. 19+ years building AI-guided, high-throughput biology platforms.
Shawn Wahi
Founding Research Engineer · Georgia Tech, Yale
Co-developed C2S-Scale at Yale. Led cardiology, oncology and aging research at Yale School of Medicine, published in Circulation and European Heart Journal.
Sravan Ravi, PhD
Founding Research Engineer · Columbia, GE HealthCare
PhD from Columbia University. Built foundation models for medical imaging as Senior AI Scientist at GE HealthCare.
Ready to predict your next program?
Give us a held-out dataset, a quantitative endpoint, and the baseline to beat. We define the evaluation before we ask you to buy anything.