Building AI software to design and test small-molecule drug candidates
Finding a molecule that binds a disease-related protein requires repeated design, synthesis and laboratory testing. Screening existing compounds explores only a small part of the possible chemical space, while promising computer designs can be difficult to make. These bottlenecks consume research budgets and slow the selection of drug candidates.
Baio Labs is building Drugon, an AI platform that helps medicinal chemists design small-molecule candidates for specific protein targets. It brings design, ordering and experimental validation into one workflow to help research teams test promising ideas faster.
What stands out about Baio Labs
In early in-vitro testing across five protein targets, more than 60% of Drugon-designed molecules showed confirmed binding.
Baio launched its first Drugon pilot projects in early 2026, with positive experimental results from industry collaborators.
Baio’s venture and research project received an EPFL AI grant as part of CHF 455,000 in combined support reported by May 2026.
Co-founder Ilia Igashov led molecular-linker design research published in Nature Machine Intelligence and co-led the RetroBridge synthesis model.
Co-founder Arne Schneuing co-led DiffSBDD, a molecular-design model published in Nature Computational Science in 2024.
Co-founder Bruno Correia built EPFL’s protein-design and immunoengineering group from 2015; the laboratory secured two ERC grants.
Spun out of EPFL to commercialize generative AI research for structure-based small-molecule drug design.
Selected for Bluelion’s 12-week Fall 2026 acceleration programme, with mentoring and investor access.
Co-founder Bruno Correia was a 2022–2023 Grass Fellow at Harvard’s Radcliffe Institute, studying computational protein design.
Received Venture Kick’s CHF 100,000 Stage 3 convertible loan, bringing cumulative programme support to CHF 150,000.
Get a curated pipeline of healthcare and life sciences startups aligned with your investment thesis.