Irina Rish

Irina Rish is a Full Professor at the Université de Montréal and a core academic member of Mila – Quebec Artificial Intelligence Institute, where she leads the Autonomous AI Lab. She holds a Canada Excellence Research Chair (CERC) and a CIFAR AI Chair, reflecting her leadership in foundational AI research. Her work is supported by major initiatives, including the U.S. Department of Energy’s INCITE project on Summit and Frontier supercomputers.

• PhD in AI, University of California, Irvine
• MSc in AI, University of California, Irvine
• MSc in Applied Mathematics, Moscow Gubkin Institute

Her research focuses on machine learning, neural scaling laws, emergent behaviors in foundation models, continual learning, robustness, and neuroscience-inspired AI. She explores how AI systems can become more general, flexible, and aligned with human cognition. Her recent work investigates training dynamics in large language models, efficient pruning techniques, and the development of time-series foundation models.

The analysis of her recent publications reveals a strong focus on scaling behaviors, continual adaptation, and robustness in AI systems. Her work spans theoretical understanding of training dynamics (e.g., zero-sum learning), practical optimization methods, and applications in climate modeling and mental health. She emphasizes open science, leading open-source projects and co-founding Nolano.ai to build efficient, compressed foundation models.

• Canada Excellence Research Chair (CERC)
• CIFAR AI Chair
• IBM Eminence & Excellence Award (2018)
• IBM Outstanding Innovation Award (2018)
• IBM Outstanding Technical Achievement Award (2017)
• IBM Research Accomplishment Award (2009)

Irina Rish advises a large group of PhD and Master’s students across Université de Montréal, McGill, and Concordia. She leads major research grants and collaborates internationally on HPC-based AI research. She is also the co-founder and CSO of Nolano.ai, driving innovation in efficient AI systems.

She leads the Autonomous AI Lab, which focuses on building large-scale foundation models, understanding neural scaling laws, and developing bio-inspired learning systems. She actively organizes reading groups on scaling, continual learning, and out-of-distribution generalization, fostering a collaborative research environment.