Join the Group

I am currently recruiting researchers for my group at Mila and Concordia University at multiple levels, including PhD students, Postdoctoral Fellows, Master's students, and PhD interns.

Research Focus

Research topics are flexible but should align with the ongoing work in our group. Potential areas of focus include:

• Large-Scale Distributed Learning: topics such as communication efficiency, decentralization, asynchrony, and model parallelism. Relevant works: ACCO, MuLoCo, A²CiD², Subnetwork Data Parallelism
• Learned Optimizers: relevant works: PyLO, μLO, Celo
• Federated Learning on the Edge: relevant work: Guiding the Last Layer in Federated Learning
• Continual Learning (especially for LLMs) and large context
• Other Topics in Efficient Training: scaling laws, distributed reinforcement learning, model distillation, neural architecture search, long-context modeling, and hyperparameter transfer.
• Mechanistic Interpretability and AI Safety

Application

Interested candidates are invited to send informal inquiries to eugene.belilovsky@concordia.ca. Please include [ReadWebsite] in your email header.