Johnson & Johnson seeks a Senior Scientist in AI Safety to establish governance frameworks and safety mechanisms for agentic AI systems in pharmaceutical R&D.; The role involves designing provable controls, conducting adversarial testing, and developing safety-native architectures to ensure autonomous workflows meet rigorous evidence standards.
Responsibilities
- Design machine-readable control policies governing agent actions, tool use, and data access in agentic workflows
- Develop and embed continuous red-teaming and adversarial evaluation methods into the GenAI platform
- Research and prototype safety-native architectures using alignment techniques and defense-in-depth approaches
- Define evidence standards, evaluation protocols,



and accountability measures for safety claims and audit trails

Requirements
- Deep expertise in AI safety, alignment research, or adversarial machine learning
- Experience designing and enforcing deterministic controls for autonomous systems
- Strong background in red-teaming, threat modeling, or AI evaluation methodologies
- Familiarity with regulated industries or quality/compliance frameworks
- Proven ability to prototype and implement safety mechanisms at system scale
- Cross-functional collaboration experience with engineering, security, and domain teams

📌 Senior Scientist (España)
🏢 Johnson & Johnson
📍 España