About

I am actively looking for research internships in academia or industry, especially positions involving large-scale training or representation learning, with access to substantial computational resources.

Hi! I’m Linxiang Su, an undergraduate student at the University of Szeged, currently working as a research assistant under the supervision of Prof. Márk Jelasity.

My current research interests include adversarial robustness, representation learning, and efficient deep learning. I am currently also engaged in a project about action understanding at RHOS lab at Shanghai Jiao Tong University. I take responsibility for large-scale data organization and cleansing, as well as million-scale training and evaluation. My work focuses on improving training efficiency and model performance, including model architecture optimization in distributed multi-GPU environments. Additionally, I designed and implemented a data caching mechanism that accelerated training and evaluation by over 10×, significantly enabling faster research iteration.

In general, I am drawn to research ideas that follow intuition and principled understanding, are grounded in theory, and are preferably simple yet extremely effective, with real-world impact.

Outside of research, I enjoy reading books on philosophy(Camus, Nietzsche etc.) and psychology(Lacan, Freud etc.), and playing table tennis. I’m also a big fan of Fan Zhendong :D