CellType is a Y Combinator W26 biotech startup building an 'agentic drug company' that simulates human biology to de-risk drug discovery. Founded by Yale professor David van Dijk and ML engineer Ivan Vrkic, it pairs Cell2Sentence, a 27-billion-parameter biological foundation model built with Google DeepMind, with AI agents that run the discovery pipeline from hypothesis to compound screening. The goal is to replace animal models and cell lines with computational simulations of human biology and cut preclinical failure rates.
Corvic AI is a Mountain View startup building an Intelligence Composition Platform - the logic layer that turns messy, multi-structured enterprise data (PDFs, tables, sensor logs, images, time series) into reliable, production-ready AI outcomes without bespoke pipeline engineering. Founded in 2023 by engineers from Intel, Katana Graph, and Determined AI, it uses proprietary techniques like Mixture of Spaces embeddings and agentic, auditable orchestration to deliver explainable enterprise intelligence for manufacturing, industrial, financial services, and life sciences customers.
Delineate is a Cambridge, Massachusetts AI company building agents that accelerate clinical trial design for pharmaceutical and biotech companies. Its custom large language models and computer vision systems extract and structure data from thousands of research papers and patents - including numbers buried inside figures and plots - turning months of manual evidence aggregation into weeks. A Y Combinator Winter 2025 company founded in 2024 by an MIT PhD and an ex-Lockheed Skunkworks AI lead, Delineate works with two top-10 pharmaceutical companies and holds an NIH STTR grant with MIT.
Salt AI is a Los Angeles contextual-AI company building a visual platform that lets regulated enterprises - especially life sciences and healthcare teams - orchestrate many AI models into governed, auditable workflows that run inside their own infrastructure. Founded in 2023 by former Myspace CTO Aber Whitcomb and Jim Benedetto, Salt lets biochemists and clinicians drag, drop, and chain models like AlphaFold2 (which Salt runs 22x faster), ProteinMPNN, and Chroma without writing code, turning months of pipeline engineering into weeks while keeping every prompt, weight, and response logged for compliance.