Post Doc - AI Safety
Johnson & Johnson MedTech Beerse, Belgium
Lead a two-year postdoctoral research program on making agentic AI safe for regulated pharmaceutical R&D, covering provable controls, adversarial assurance, and safety-native architecture. You will design reproducible experiments, prototype layered architectures, and work with platform engineers to move validated methods into a GenAI platform. The role is hands-on research with real systems as the testbed, not frontier model pre-training or production engineering.