Mido Assran

Mahmoud (Mido) Assran is a Research Scientist working on world models and self-supervised learning, with a focus on Joint-Embedding Predictive Architectures (JEPA), a framework for learning to understand the world by predicting "abstractions" rather than raw data, enabling goal-driven agents to act efficiently and safely in the real world. Mido completed his doctoral studies at McGill University and Mila, the Quebec AI Institute.