Director, AI and Data Science

Role Summary 

Lexeo is at an inflection point where AI and advanced analytics can materially accelerate decision-making across discovery, development, and operational execution. This Sr. Director will set direction and deliver applied AI/ML solutions across internal workflows and externally facing outputs, ranging from R&D insights to partner-ready analyses, while partnering closely with scientific teams and, when needed, external vendors/partners to solve real problems. This role is intentionally hands-on and outcome-driven: a leader who can build, validate, and operationalize models using real-world biopharma data to raise the signal-to-noise ratio in small or unstructured datasets (including synthetic control arm approaches where appropriate). 

Key Responsibilities

AI/ML Strategy + Delivery  

• Define and execute Lexeo’s applied AI/ML roadmap across discovery and development, prioritizing use cases that improve speed, quality, and decision confidence. 
• Deliver solutions that are internal-only (e.g., scientific decision support, operational forecasting) and those that are generated internally but external-facing (e.g., partner-ready analyses (regulatory dossiers, briefing books, protocols etc.), validated dashboards, and decision materials).
• Establish best practices for model lifecycle management (validation, documentation, monitoring, retraining), especially where outputs influence scientific decisions or regulated workflows. 

Advanced Analytics + Predictive Modeling 

• Lead development and selection of appropriate ML approaches (e.g., XGBoost, Random Forest, SVMs, and other advanced models) based on problem framing, data constraints, interpretability needs, and deployment context.
• Build and oversee predictive analytics using real-world data, including robust evaluation design, bias/variance trade-offs, and performance monitoring. 

Small Data Excellence + Synthetic Controls  

• Apply techniques to amplify signal-to-noise in smaller datasets (e.g., regularization, Bayesian methods, hierarchical modeling, augmentation, multimodal integration, careful feature engineering, uncertainty quantification).
• Guide strategy for synthetic control arms and comparable approaches (as appropriate), ensuring methodological rigor, transparency, and fit-for-purpose use in decision-making. 

Drug Discovery / Translational Partnership 

• Translate drug discovery and translational questions into testable analytical hypotheses; partner with bench scientists to design data capture that enables strong modeling.
• Serve as a bridge between scientific teams and data/engineering, ensuring solutions are scientifically credible and operationally adoptable.