Google DeepMind alumni unveil Bioptimus: Aiming to build first universal biology AI model


French startup ecosystem sees rapid growth, with companies like Bioptimus emerging as key players. Bioptimus, based in Paris, raises $35 million in seed funding to develop the first universal AI model for biology. The startup aims to connect various levels of biology using generative AI, from molecules to organisms.

The team at Bioptimus includes former Google DeepMind and Owkin scientists, leveraging AWS computing and Owkin’s data generation capabilities. By collaborating with academic hospitals globally, Bioptimus has access to extensive multimodal patient data for model training.


Bioptimus’s CEO, Jean-Philippe Vert, highlights the company’s agility compared to larger tech firms, allowing for faster data access and collaboration with partners. The team also emphasizes transparency and open-source initiatives, similar to Mistral’s model releases.


Currently, AI models in biology focus on specific aspects like protein sequences or cell images. Bioptimus aims to create a unified model that captures the entirety of biology, leveraging converging AI technologies.


Vert stresses that data access is a critical challenge, which Bioptimus addresses through its partnership with Owkin. The company’s emphasis on collaboration and trust-building sets it apart in the AI biotech landscape.



Prompt Engineering Guides



©2024 The Horizon