Nucleate Podcast How Turbine Accelerates Drug Discovery with AI Simulation | Szabi Nagy, CEO and Co-Founder of Turbine
13 snips
Feb 24, 2026 Szabi Nagy, CEO and co-founder of Turbine and former Tresorit founder, pivoted from economics to build AI-driven virtual biology. He discusses virtual experiments that speed drug discovery, the three phases of applying their models, forming pharma partnerships, fundraising in tight markets, and finding resilient mentorship and team structures.
AI Snips
Chapters
Transcript
Episode notes
Three Phases Where Simulations Add Value
- Turbine's platform applies across target ID, hit-to-lead optimization, and translational patient selection.
- They simulate target screens, prioritize programs and biomarkers, and aim to predict trial design from early clinical data.
Design Models To Generalize Not Overfit
- Build models to generalize across diseases and cell types instead of narrowly optimizing for one assay.
- Combine public, partner, and targeted in-house seed data and harmonize heterogeneous snapshots to enable out-of-distribution predictions.
Multimodal Seed Data And A Virtual Lab
- Turbine ingests multimodal pre/post perturbation data from cells, organoids, PDXs and patients and uses targeted in-house assays as seed data.
- They built a virtual lab interface so biologists can set up predictive models and automatically launch validations.