At Lilly, the business problem is to accelerate drug discovery by developing continuously improving AI models that benefit both Lilly and its biotech partners. The TuneLab platform aims to solve this by providing biotech companies with access to machine learning models trained on Lilly's proprietary pharmaceutical research data, enabling collaborative drug discovery while preserving data privacy.
Requirements
- Strong experience with molecular property prediction and QSAR/QSPR methods
- Deep understanding of medicinal chemistry principles and ADMET optimization
- Experience with federated learning and distributed optimization in chemical applications
- Expertise in graph neural networks and geometric deep learning for molecules
- Strong background in organic chemistry and synthetic feasibility assessment
- Experience with fragment-based and structure-based drug design
- Proficiency in cheminformatics tools (RDKit, DeepChem)
Responsibilities
- Architect and implement advanced multi-task learning models specifically for small molecule properties including ADMET endpoints, solubility, permeability, metabolic stability, and off-target liabilities, handling diverse chemical representations (SMILES, graphs, 3D conformations).
- Design and deploy state-of-the-art generative models (VAEs, diffusion models, flow matching, autoregressive models) for de novo small molecule design, lead optimization, and scaffold hopping that respect synthetic accessibility and drug-likeness constraints.
- Develop integrated prediction-generation pipelines that optimize molecules simultaneously across multiple ADMET properties while maintaining target potency, using techniques like multi-objective optimization and Pareto front exploration.
- Implement algorithms for efficient exploration of synthetically accessible chemical space, including reaction-aware generation, retrosynthetic planning integration, and fragment-based design approaches.
- Build models that learn and exploit structure-activity relationships from sparse, noisy bioactivity data across federated partners, including matched molecular pair analysis and activity cliff prediction.
- Develop self-supervised and semi-supervised methods to learn robust molecular representations from large collections of unlabeled compounds, enabling better generalization to novel chemical series.
- Create AI-driven workflows for common medicinal chemistry tasks including bioisosteric replacement, metabolic site prediction, toxicophore removal, and property optimization while maintaining intellectual property considerations.
Other
- PhD in Computational Chemistry, Cheminformatics, Medicinal Chemistry, Chemical Engineering, or related field from an accredited college or university
- Minimum of 2 years of experience in small molecule drug discovery
- up to 10% travel (attendance expected at key industry conferences).
- Relocation is provided.
- Understanding of IP considerations in generative molecular design
