Hi! I'm Hiroki Naganuma, a Ph.D. candidate in Computer Science at Université de Montréal and Mila - Quebec AI Institute, advised by Prof. Ioannis Mitliagkas. My research focuses on large-scale parallelization in machine learning and the analysis of training dynamics in deep neural networks. I recently completed student researcher role at Google DeepMind (Mountain View), where I worked under the guidance of Dr. George E. Dahl on learning rate scheduling. Previously, I completed a summer internship at Microsoft Research (Redmond), where I worked on efficient pretraining algorithms for large language models with Dr. Philipp Witte and Dr. Russell J. Hewett. In the summer of 2025, I will be a research intern at Meta - AI and Systems Co-Design (Menlo Park) (host: Dr. Hao-Jun Michael Shi and Dr. Parameswaran Raman).
I am a recipient of the Masason Foundation Fellowship. I earned my B.Sc. (2017) and M.Sc. (2019) from the Tokyo Institute of Technology as Valedictorian, where I had the privilege of working closely with Prof. Rio Yokota and an outstanding group of collaborators.
Collaboration & Mentorship: If you're interested in my work, feel free to reach out!
📄 CV (Last updated: April 2025)