Summary
Xin Zhang is a PhD candidate in Data Science & Analytics at HKUST with eight years of experience building distributed ML systems and large-scale data platforms. He has practical intern and research experience at Microsoft, Baidu, Tencent, and Peking University, contributing to distributed training frameworks (including pipeline-parallelism strategies like GPipe and Pipedream) and production-scale CTR model training for ~1TB models. His work spans MLSys, GNNs, and AIOps—developing anomaly detection and root-cause pipelines for large clusters—demonstrating a blend of research rigor and production engineering. Based in Hong Kong, he is an active open-source contributor (see xinzhang.me) who prefers solving hard systems problems that enable training extremely large models across multi-GPU clusters.
8 years of coding experience