about me

I am an incoming PhD student at EPFL in Electrical Engineering where I will join Florent Krzakala’s lab IdePHICS in Lausanne, Switzerland. Previously, I had the great opportunity to work with Murat A. Erdogdu at the Vector Institute for a research internship. Before that, I am grateful to have completed my Master’s thesis at ETH Zürich with Sara van de Geer.

email: jivan [dot] waber [at] gmail [dot] com

Currently in Lausanne, Switzerland

research interests

I am interested in discovering the underlying mechanisms of artificial intelligence and get insights about learning and intelligence. To do so, I use theoretical tools and computational experiments to better understand machine learning phenomena. At the moment, I am particularly interested in learning dynamics and scaling laws.

news

[July 2025] Presented a paper about the research I did with colleagues at the Vector Institute at the 3rd High-dimensional Learning Dynamics workshop at ICML 2025 (Oral).

[September 2024] The work I conducted with my colleagues at MeteoSwiss was presented at the European Meteorological Society Annual Meeting in Barcelona (here’s the abstract).

[April 2024] Started a research internship at the Vector Institute in Toronto with Prof. Murat A. Erdogdu on deep learning theory.

[February 2024] Graduated from ETH Zürich with a MSc Applied Mathematics with application in machine learning.

[October 2023] Started developing deep learning models for thunderstorm prediction for my mandatory civil service at MeteoSwiss, the Federal Office for Meteorology and Climate in Zürich.

[August 2023] Defended my Master’s thesis on the topic of benign overfitting supervised by Prof. Sara van de Geer and Dr. Geoffrey Chinot.