Postdoctoral Scientist – Multimodal Representation Learning for Predictive Biology

About the role

Johnson & Johnson Innovative Medicine is recruiting a Postdoctoral Scientist to join its Data, Data Science & Artificial Intelligence (DDSAI) organization. The role works at the intersection of advanced AI/ML modeling, multimodal representation learning, drug discovery and predictive biology, developing models that analyze and quantify the heterogeneity of perturbed cellular systems and integrate multi-scale biological data such as high-content imaging, phenomics, transcriptomics and proteomics. It is a two-year fixed-term position based in Madrid, Spain, with the same role also open in Cambridge (MA), Spring House (PA) and Beerse (Belgium).

Key responsibilities

• Design, develop, implement and validate AI/ML solutions for drug discovery using multimodal data of perturbed cells.
• Analyze large-scale, heterogeneous high-dimensional data (imaging microscopy, phenomics, transcriptomics, proteomics) to extract biological insights.
• Develop and evaluate computer vision solutions to quantify cell heterogeneity.
• Collaborate with internal and external scientists and engineers to advance algorithms and product development.
• Communicate complex technical methods and present findings to stakeholders.
• Maintain documentation and an up-to-date code repository.
• Draft manuscripts and share research findings in peer-reviewed conferences and journals.

Required profile

• Ph.D. in Electrical Engineering, Biomedical Engineering, Computer Science, Applied Mathematics or a related field, completed within the last 3 years or to be completed within the next 6 months.
• Proven track record of publications in top-tier AI/ML conferences or journals.
• Experience developing foundation models or multimodal modeling of biological data, such as Cell Painting, imaging microscopy and multi-omics.
• Strong communication skills and ability to present results in a multi-functional setting.

Required skills

• Solid knowledge of deep-learning architectures (Transformers, CNNs, graph networks, self-supervised and multi-instance learning) applied to multimodal representation learning, alignment and fusion.
• Proficiency in one or more programming languages, preferably Python, and in AI frameworks such as PyTorch or TensorFlow.
• Expertise in at least one domain: representation learning, multimodal integration, computational biology or imaging microscopy.

What we offer

• Base pay range of EUR 43,600 to EUR 70,150 for positions hired in the United States; benefits vary by location.
• Annual bonus, vacation days, parental leave of at least 12 weeks, bereavement and caregiver leave, well-being reimbursement and health programs.