Postdoc: Safe AI Architectures for Pharma R&D

7555-Janssen Pharmaceutica NV Legal Entity; Geel; Hybrid; EUR 52,000 - 75,000; Full time

Benefits offered by this job

Annual bonus

Vacation days

Parental leave

Bereavement leave

Well-being reimbursement

Job summary

Johnson & Johnson Innovative Medicine is recruiting a Postdoctoral Researcher, AI Safety, to join the Data, Data Science & AI organization. The role is a two-year fixed-term appointment reporting to the Scientific Fellow, AI Safety, with base options at Beerse, Belgium or other sites in Europe and the UK.

You will lead a focused research program on governance of agentic AI in regulated R&D, ensuring safety by construction.

Qualifications

• PhD in computer science, AI/ML, applied mathematics, or closely related field.
• First-author publications or preprints in AI safety/alignment or related areas.
• Hands-on experience with foundation models and agentic AI.
• Strong programming skills and reproducible experiments.

Responsibilities

• Research program: execute an independent agenda on controls, adversarial assurance, or safety-native architecture.
• Design rigorous, reproducible experiments with baselines, ablations, and uncertainty estimates.
• Evaluate pre-training data interventions and post-training methods for safety properties.
• Prototype layered architectures to constrain agent behavior and translate requirements.
• Collaborate with platform engineering to move methods from prototype to GenAI Platform.
• Present work to scientific, engineering, and leadership audiences; represent at conferences.

Skills

PhD in CS/AI/Math

First-author publications

AI safety

Foundation models

Agentic AI

Python

ML frameworks

English proficiency

Model evaluation

Education

PhD in computer science, AI/ML, applied mathematics

Tools

Python

ML frameworks

Job description

Johnson & Johnson Innovative Medicine is recruiting a Postdoctoral Researcher, AI Safety, to join the Data, Data Science & AI organization. The role is a two-year fixed-term appointment reporting to the Scientific Fellow, AI Safety, with base options at Beerse, Belgium or other sites in Europe and the UK.

You will lead a focused research program on governance of agentic AI in regulated R&D, ensuring safety by construction.