CellType

CellType is building foundation models and agent systems for biology.

We are building models that learn from biological data, experiments, and outcomes directly. Our goal is to develop AI systems that are genuinely useful for scientific discovery and drug development.

We work on problems such as preclinical-to-clinical translation, response prediction, biomarker discovery, and reasoning across complex biological datasets. These are technically difficult problems with real scientific and commercial importance.

We are an early team with high standards for technical depth, speed, and ambition. This is a good fit for people who want to work on hard model and systems problems in a domain that matters.

We are building the infrastructure needed to train and deploy foundation models for biology.

This includes petabyte-scale data pipelines, indexing and retrieval systems over large biological datasets, and multi-node GPU training infrastructure for model training, post-training, and evaluation. We care about the full path from raw data ingestion to reproducible datasets to efficient distributed training and reliable production systems.

Key technical areas include:

• large-scale biological data ingestion and processing
• retrieval, indexing, and search over heterogeneous experimental and molecular datasets
• dataset validation, reproducibility, and versioning
• distributed data loading and performance optimization
• multi-node GPU training and post-training infrastructure
• observability, debugging, and reliability for ML systems
• internal platforms and tooling for fast research iteration

We are interested in people who are strong in distributed systems, data infrastructure, ML systems, or large-scale training, and who want to apply that skillset to biology.