About the job
The Computational Toxicology Group within Nonclinical Drug Safety (NDS) seeks a senior AI/ML scientist to drive the development and deployment of next-generation computational toxicology capabilities. This role will combine advanced machine learning, foundation model engineering, and domain expertise to accelerate safer drug discovery and support regulatory-ready New Approach Methodologies (NAMs). The successful candidate will lead cross-functional projects, deliver production-grade models and agentic systems, and help establish governance and MLOps practices that ensure reproducibility, transparency, and ethical AI use in preclinical research.
Responsibilities
Lead deployment of advanced AI/ML solutions (multimodal transformers, graph or sequence models, Bayesian/probabilistic approaches) for toxicity prediction and translational safety applications.
Design and implement agentic AI systems tailored to toxicology use cases
Specialize in the fine-tuning and alignment of foundation models for toxicology domain-specific applications and supporting new approach methods (NAMs).
Drive collaboration with cross-functional teams of toxicologists, computational scientists, biologists, and chemists to ensure explainability, reproducibility, and address specific "context of use" regulatory requirements for safety assessments.
Champion best practices in model governance, and responsible AI within a regulated environment, helping to establish frameworks for responsible and ethical AI deployment in preclinical research.
Present and communicate science in key internal and external toxicology forums.
Qualifications
Minimum
Ph.D. or M.S. in Computer Science, Computational Biology, Computational Chemistry, Bioinformatics, Statistics, or related field.
0+ years post-PhD or 3+ years post-MS experience developing and deploying AI/ML models
Hands-on experience with large language models and agentic AI frameworks (fine-tuning, prompt engineering, multi-agent orchestration, tool use, and API-based production orchestration) required.
Proven experience integrating and modeling multimodal datasets (omics, chemical, textual, imaging).
Strong software development skills in Python and familiarity with modern ML frameworks (e.g., PyTorch, TensorFlow), MLOps tools, cloud platforms (AWS preferred), and HPC environments.
Excellent communication skills; ability to translate complex technical work to domain experts and leadership.
Preferred
Demonstrated publication record applying AI/ML to life sciences or toxicology.
Experience with probabilistic/Bayesian modeling, uncertainty quantification, or causal inference.
Prior experience designing agentic systems, human-in-the-loop workflows, or using reinforcement learning for agent behavior control.
Prior experience working in regulated environments or developing regulator-ready models.