Director, World Model & Agentic Learning

Role: Lead the AI science team focused on developing J&J's enterprise world model and agentic-learning capabilities within the Generative AI organization. Primary objectives include designing systems for knowledge accumulation, reasoning, self-assessment, and continuous improvement from real-world operation and expert feedback. Responsibilities encompass setting technical direction, building mechanisms for knowledge representation, confidence estimation, and reasoning, as well as partnering with scientists and domain experts to embed expert judgment into AI systems. The role involves leading a team of 4–8 AI scientists, establishing a culture of rigor and accountability, and collaborating across R&D, data, and AI evaluation functions. Qualifications: minimum 8 years in industry AI/ML system development, with hands-on experience in large language models, knowledge representation, and systems emphasizing continual learning and knowledge accumulation. Preferred: advanced degree (PhD), experience in regulated environments such as life sciences or healthcare, and expertise in knowledge graphs or explicit knowledge systems. High-value: focus on knowledge-based reasoning, continual learning, agentic AI, and application within life sciences or healthcare R&D. Work setup: hybrid locations in the US, including Cambridge, MA; La Jolla, CA; Titusville, NJ; or Spring House, PA.