Hao Du

I recently joined the AI Infra group at TeleAI, the Institute of Artificial Intelligence, China Telecom, as a Researcher. I focus on generative AI infrastructure, particularly systems for diffusion models and world models. Previously, I worked at Huawei 2012 Lab on inference optimization and in-network computing, and held research internships at Microsoft Research Asia, Momenta, and SenseTime.

I received my Ph.D. in Computer Science from City University of Hong Kong in 2025, advised by Prof. Jing Liao at City University of Hong Kong and Prof. Yan Xu at Beihang University. I earned my B.S. in Software Engineering from the University of Electronic Science and Technology of China in 2019.

My research focuses on efficient and scalable infrastructure for generative models, spanning language models, diffusion models, and world models. Building on my experience in sparse attention, KV-cache systems, and collective communication, I aim to bring model capabilities into practical systems with lower computational cost and latency. I am currently working on infrastructure acceleration for Telefuser; this project is ongoing.