Senior Principal / Associate Director, Scientific ML for Drug Discovery

Lila Sciences•Cambridge, MA

54d•$228,000 - $480,000

About The Position

Your Impact at Lila Lead and scale a cross-functional Scientific ML team that delivers end-to-end impact on real programs. You will be the player–coach setting technical direction across AI structure-based design, ligand-based optimization, synthesis planning, ADMET/PK modeling, and AI-accelerated physics, while partnering with ML platform engineering to ship reliable, production-grade services. Your leadership will turn diverse data and models into a cohesive, closed-loop design engine that shortens DMTA cycles, improves hit rate and MPO, and de-risks program decisions. What You'll Be Building Strategy and roadmap: Define the technical vision and quarterly milestones for SBDD, ligand-based QSAR/ADMET, synthesis planning, and physics-ML; prioritize along live program needs and compute budget. Team building: Hire, mentor, and develop a 6+ person team spanning AI scientists and an ML platform engineer; establish high standards for scientific rigor, code quality, and collaboration. Unified design loop: Orchestrate a synthesis-aware, MPO-constrained, uncertainty-calibrated design workflow that fuses assay-driven ligand models with structure/physics signals and ADMET/PK constraints. Evaluation governance: Institute leakage-safe datasets and splits (scaffold/time/series), prospective validations, OOD tests, and model gating; publish model cards and decision logs for auditability. Data contracts and foundations: Co-design schemas, ontologies, and provenance with Assay Informatics, Structural Biology, and Data Platform; ensure reliable ETL from ELN/LIMS, structure, and simulation. Productionization: Partner with ML Engineering to deliver reproducible training, scalable serving (APIs/batch), monitoring, and incident response for scientific services on cloud + HPC. External collaboration: Coordinate with partner teams internal and exteral to Lila for assay QC, structural prep, and data platform SLAs; evaluate vendors and open-source where it accelerates impact. Culture and communication: Set a high bar for clarity, integrity, and humility; communicate uncertainty and trade-offs to technical and executive stakeholders.

Requirements

8+ years (post-PhD or equivalent) building and shipping ML for drug discovery or closely related domains; demonstrated impact on live programs
Technical depth and breadth: Expertise in at least two of the following and fluency across the rest: AI SBDD (equivariant/3D graph models for pose/affinity, pocket embeddings) Ligand-based QSAR/ADMET and active learning for hit-to-lead/lead opt Synthesis planning and reaction/condition/yield modeling ADMET/PK/PD (IVIVE, PBPK/QSP) and uncertainty/calibration ML-for-simulation/free energy (Δ-learning surrogates, learned force fields)
ML engineering excellence: PyTorch/JAX, geometric learning, generative modeling, experiment tracking, model/data versioning, serving; comfort with hybrid cloud + HPC.
Scientific rigor: Statistical mechanics and thermodynamics basics, medicinal chemistry and DMPK fundamentals, assay QC and leakage control; designs prospective, decision-grade evaluations.
Leadership: Hires and grows high-performing teams; sets crisp priorities; aligns diverse stakeholders; communicates clearly at both the whiteboard and the exec table.

Nice To Haves

PhD in CS, Computational Chemistry, Chemoinformatics, Biophysics, or related field with publications in top ML/drug discovery venues.
Delivered unified design loops that improved hit rate/MPO and reduced cycle time; experience integrating retrosynthesis and PBPK into optimization.
Open-source leadership (e.g., RDKit/Chemprop/DeepChem, PyTorch Geometric/e3nn, OpenMM) or vendor evaluation/deployment experience.
Experience with HTS/DEL analytics, structural bioinformatics (AlphaFold/ensembles), or regulated documentation (model qualification).

Responsibilities

Define the technical vision and quarterly milestones for SBDD, ligand-based QSAR/ADMET, synthesis planning, and physics-ML; prioritize along live program needs and compute budget.
Hire, mentor, and develop a 6+ person team spanning AI scientists and an ML platform engineer; establish high standards for scientific rigor, code quality, and collaboration.
Orchestrate a synthesis-aware, MPO-constrained, uncertainty-calibrated design workflow that fuses assay-driven ligand models with structure/physics signals and ADMET/PK constraints.
Institute leakage-safe datasets and splits (scaffold/time/series), prospective validations, OOD tests, and model gating; publish model cards and decision logs for auditability.
Co-design schemas, ontologies, and provenance with Assay Informatics, Structural Biology, and Data Platform; ensure reliable ETL from ELN/LIMS, structure, and simulation.
Partner with ML Engineering to deliver reproducible training, scalable serving (APIs/batch), monitoring, and incident response for scientific services on cloud + HPC.
Coordinate with partner teams internal and exteral to Lila for assay QC, structural prep, and data platform SLAs; evaluate vendors and open-source where it accelerates impact.
Set a high bar for clarity, integrity, and humility; communicate uncertainty and trade-offs to technical and executive stakeholders.