Experteer Overview In this role, you will advance AI safety in pharmaceutical Ru0026amp;D by defining provable controls, adversarial assurance, and safety-native architectures for agentic AI.

You will work within the Data, Data Science u0026amp; AI organization in Madrid or Barcelona, partnering with cross-functional teams to translate safety requirements into scalable platform safeguards.

Your work shapes governance and verification evidence at scale, contributing to safe and trusted AI-enabled workflows.

This is an impactful, hands-on scientific position with visibility to regulators and stakeholders, offering a path to influence next-generation AI in life sciences.

Compensaciones / Ventajas• Design machine-readable control policies governing agent actions, data access, and information flow across workflows• Implement deterministic policy enforcement with auditable decisions and human-approval paths• Enable owners to author, test, and maintain controls via accessible policy interfaces• Develop continuous red-teaming methods for agentic AI and integrate them into the GenAI Platform• Define evaluation protocols and evidence thresholds to distinguish safety properties from unverified claims• Research safety-native architectures and data interventions across training stages (pre- and post-training)• Translate regulatory, quality,



and privacy requirements into testable system specifications• Define accountability measures and ensure audit-reconstructable evidence for safety claims• Collaborate with Ru0026D, platform engineering, and security/privacy/legal/quality teams to shape scientific direction and external contributions

Responsabilidades• PhD in computer science, AI/ML, applied mathematics, or related field (required)• At least 1 year of post-PhD research or industry experience in AI/ML, autonomous agents, or security-critical systems• Deep hands-on expertise in agentic AI, foundation models, retrieval-augmented generation, tool orchestration, memory, planning, multi-agent frameworks, and their failure modes• Proven expertise in AI safety and alignment (e.g., supervised fine-tuning, preference optimization, safety tuning, adversarial evaluation, interpretability, scalable oversight)• Experience in policy-as-code and authorization; adversarial ML and AI red teaming; information-flow or data-access control; provenance and lineage; formal specification or verification• Strong AI engineering skills: prototyping, reproducible experiments, production-ready code• Excellent written and verbal communication; ability to present safety arguments to scientific, engineering, and executive stakeholders• Scientific rigor in characterizing model/agent behavior, uncertainty, limitations, and evidence strength

Requisitos principales• hybrid work model• annual bonus based on performance• vacation days• minimum 12 weeks parental leave• well-being reimbursement• insurance plans (variable by location)

📌 Senior Scientist - AI Safety (Barcelona)
🏢 Johnson u0026 Johnson
📍 Barcelona