The Data Pulse Building a bilingual culture for data-first drug discovery with Daphne Koller (insitro)
12 snips
Jul 28, 2020 
Daphne Koller, Founder and CEO of Insitro, is a pioneer in merging machine learning with biomedicine. In the discussion, she emphasizes the importance of a bilingual culture that bridges biology and ML for innovative drug discovery. Koller shares insights on building hybrid teams, the challenges of large biomedical datasets, and how ML can enhance drug efficacy predictions. She also highlights successful strategies for managing clinical trials and encourages integrating data scientists into early decision-making in pharma. A must-listen for anyone interested in the future of healthcare!
AI Snips
Chapters
Transcript
Episode notes
Marry Fast Learning With Biological Patience
- Tech's rapid cycles (MVPs, fast failure) can inform biotech but must respect biological pace.
- Insitro aims for intermediate cadence using automation and ML in experimental loops.
Insitro's Engineered Hybrid Approach
- Insitro is a hybrid: robotic, reproducible experiments integrated with ML in the inner loop.
- That hybrid enables repeatable experiments at scale suited to model-guided discovery.
Target Decisions Drive Clinical Failures
- Drug R&D is a long 15-year path with many forked decisions often driven by heuristics.
- ML can make better, engineered predictions at key decision points to reduce failure.