Multiple PhD Positions in Foundation Models for Scientific Discovery

• JEPA-like Architectures
• Representation Learning

We are opening multiple fully funded PhD positions within the FUNDIS (Foundation Models for Scientific Discovery) project at the University of Geneva .

FUNDIS is a multidisciplinary research initiative developing the next generation of AI methods for scientific discovery by combining modern foundation models with domain sciences.

Research topics include

• Foundation Models
• Self-Supervised Learning
• JEPA-like Architectures
• World Models
• Representation Learning
• Explainable AI
• Information Bottleneck Theory
• Bio-inspired AI
• Reasoning and Memory
• Vision-Language Models
• Scientific Machine Learning

Scientific application domains

• Astronomy (SKA, JWST, Euclid)
• High-Energy Physics (ATLAS/CERN)
• Climate Science
• Computational Art History & Digital Humanities
• Global Governance and Computational Diplomacy

What We Offer

• Fully funded PhD positions with competitive Swiss salaries
• An outstanding multidisciplinary research environment
• Close collaboration with internationally recognized researchers
• Regular scientific seminars, reading groups and interdisciplinary discussions
• Opportunities to publish at leading AI and scientific conferences and journals
• Access to scientific datasets
• Research at the intersection of AI, mathematics, physics, astronomy, climate science, digital humanities, and computational social sciences

We welcome highly motivated applicants with backgrounds in Computer Science, Artificial Intelligence, Mathematics, Physics, Astronomy, Cosmology, Statistics, Art, Digital Humanities, Electrical Engineering, Data Science, or closely related fields.

Application deadline: 30 August 2026, 23:59 CET.