There is a new startup utilizing generative AI technology located in Paris. What sets Bioptimus apart is its unique focus on applying the collective knowledge gained from AI models specifically to the field of biology.
The decision to concentrate exclusively on biology stems from the complexity of obtaining training data in this domain. While OpenAI is transitioning towards licensing agreements with content providers instead of web crawling, Bioptimus faces distinct challenges due to the sensitive nature of clinical data, which is not readily accessible to the public.
Similar to other AI ventures, Bioptimus will require significant capital investment for training its models using high-performance GPUs and recruiting skilled researchers. To support these endeavors, the startup has secured a $35 million seed funding round led by Sofinnova Partners, with participation from Bpifrance’s Large Venture fund, Frst, Cathay Innovation, Headline, Hummingbird, NJF Capital, Owkin, Top Harvest Capital, and Xavier Niel.
Bioptimus is not emerging out of obscurity. Jean-Philippe Vert, serving as the co-founder and executive chairman in a non-operational capacity, brings his expertise from his role as the Chief R&D Officer at Owkin, a prominent French biotech company focused on leveraging AI for drug discovery and diagnostics.
Rodolphe Jenatton, Bioptimus’ CTO, boasts extensive experience in artificial intelligence, having previously held a senior research scientist position at Google. Additionally, several co-founders have backgrounds as former researchers at Google DeepMind.
Image Credits: Bioptimus
Through partnerships with renowned academic hospitals worldwide, Owkin, in its collaborations with top biopharmaceutical companies, has accumulated diverse patient data sets. Bioptimus intends to utilize this rich data pool to train its foundational model effectively.
A Revolutionary Initiative by Owkin
Bioptimus could be viewed as a spin-off or a visionary project stemming from Owkin. The decision to establish a separate entity for developing foundational models was driven by the immense complexity involved in creating new AI models.
Jean-Philippe Vert explained to TechCrunch, “Building biology foundational models is not part of Owkin’s roadmap, but Owkin is supportive and eager to collaborate with a company like Bioptimus. Training very large-scale foundational models necessitates significant resources in terms of data volume, computing power, and data modalities, which are more feasible to access as a dedicated entity.” As a specialist in foundational models, Bioptimus is better equipped to undertake this ambitious task.
The startup has also forged a partnership with Amazon Web Services, indicating that its model will be trained in Amazon’s advanced data centers. With substantial funding in place, Bioptimus is now poised to develop its AI model and explore the possibilities it offers to the biotech research community.
Vert envisions, “The AI we develop will enhance disease diagnosis, precision medicine, and facilitate the creation of new biomolecules for medical or environmental applications.”
