From Hype to Practice: Where AI Actually Helps in Drug Discovery

What You Will Learn

• Identify realistic use cases for ML/AI in small-molecule drug discovery, and distinguish them from “solution in search of a problem” projects.
• Describe how data volume, assay design, and biological complexity limit model performance and what can be done about it.
• Outline a pragmatic approach to organizing assay and registration data (e.g., in CDD Vault) to support future ML efforts, including the role of ontologies and standardized metadata.
• Understand emerging tools and approaches (foundation models, generative models with synthetic constraints, DEL structure–function mapping) and how they might integrate into your pipeline.
• Articulate different strategies for balancing openness (methods, challenges, shared data) with IP protection around actual chemical matter.
• Apply practical career and team-building advice to recruit, train, or become scientists who are comfortable working at the interface of biology, chemistry, and computation.