Zhipeng Zhang (张智鹏)

Beijing, China Email GitHub Google Scholar

I am a software engineer at Alibaba, working on pre-training and post-training infrastructure for LLMs. I am a core contributor to the pre-training infrastructure that powers Qwen2.5, Qwen3, and Qwen3.5. My recent work focuses on Agentic RL infrastructure for the Qwen 3.5/3.6 series. Previously, I co-founded Flink ML, a distributed machine learning framework built on Apache Flink, where I serve as a Flink Committer.

I received my Ph.D. in Computer Science and Technology from Peking University in 2020, advised by Prof. Bin Cui. My doctoral research focused on big data systems and distributed machine learning systems. I earned my B.S. from Shandong University (Taishan College) in 2015, advised by Prof. Xiaohui Yu.

Work Experience

  • Staff Engineer, Alibaba, Beijing. (07/2020 - Present)
  • Research Intern, Tencent, Beijing. (11/2018 - 11/2019)
  • Visiting Researcher, ETH Zurich, Switzerland. (07/2017 - 01/2018)

Publications

  • Accelerating Compound LLM Training Workloads with Maestro

    Xiulong Yuan, Hongqing Chen, Jiaxuan Peng, Fan Zhou, Zhixiang Ruan, Zekun Wang, Bo Zheng, Rui Men, Haiquan Wang, Zhipeng Zhang, Langshi Chen, Man Yuan, Jiaqi Gao, Zhengping Qian, Junyang Lin, Yong Li, Wei Lin, Junhua Wang, Jingren Zhou

    arXiv, 2026 (arXiv)

  • VRouter: Micro-batch Level Load Balance via Inter-EP Routing for MoE Training

    Haiquan Wang, Zhipeng Zhang, Guanshujie Fu, Youhui Bai, Jiangfei Duan, Yuan Man, Langshi Chen, Hongqing Chen, Siyu Wang, Xiulong Yuan, Yunfei Mao, Si Chang, Linlang Jiang, Yingtao Li, Yan Wang, Yong Li, Wei Lin, Cheng Li

    Preprint, 2026

  • AdaHC: Accelerating Multi-Token Prediction with Adaptive Head Chunking with Pipeline Parallelism

    Yan Wang, Chang Si, Kaiming Yang, Zhipeng Zhang, Weijian Liu, Man Yuan, Mingzhen Li, Yong Li

    ICML, 2026

  • TrainMover: An Interruption-Resilient Runtime for ML Training

    ChonLam Lao, Jiaqi Gao, Jiamin Cao, Zhipeng Zhang, Pengcheng Zhang, Jiangfei Duan, Minlan Yu, Aditya Akella, Zhilong Zheng, Yu Guan, Yichi Xu, Yong Li, Ennan Zhai, Dennis Cai, Zhengping Qian, Jingren Zhou

    OSDI, 2026

  • A Few GPUs, A Whole Lotta Scale: Faithful LLM Training Emulation with PrismLLM

    Shaoke Xi, ChonLam Lao, Boyi Jia, Jiaqi Gao, Zhipeng Zhang, Jiamin Cao, Brian Sutioso, Erci Xu, Minlan Yu, Kui Ren, Yong Li, Zhengping Qian, Ennan Zhai, Jingren Zhou

    SOSP, 2026 (arXiv)

  • Qwen3 Technical Report

    Qwen Team (LLM infra contributor)

    arXiv, 2025 (arXiv)

  • Qwen2.5 Technical Report

    Qwen Team (LLM infra contributor)

    arXiv, 2024 (arXiv)

  • Infinite-LLM: Efficient LLM Service for Long Context with DistAttention and Distributed KVCache

    Bin Lin, Chen Zhang, Tao Peng, Hanyu Zhao, Wencong Xiao, Minmin Sun, Anmin Liu, Zhipeng Zhang, Lanbo Li, Xiafei Qiu, Shen Li, Zhigang Ji, Tao Xie, Yong Li, Wei Lin

    arXiv, 2024 (arXiv)

  • Model Averaging in Distributed Machine Learning: A Case Study with Apache Spark

    Yunyan Guo, Zhipeng Zhang, Wentao Wu, Jiawei Jiang, Ce Zhang, Bin Cui, Jianzhong Li

    VLDBJ, 2021

  • ColumnSGD: A Column-oriented Framework for Distributed Stochastic Gradient Descent

    Zhipeng Zhang, Wentao Wu, Jiawei Jiang, Lele Yu, Bin Cui, Ce Zhang

    ICDE, 2020

  • PSGraph: How Tencent trains large-scale graphs with Spark?

    Jiawei Jiang, Pin Xiao, Lele Yu, Xiaosen Li, Jiefeng Cheng, Xupeng Miao, Zhipeng Zhang, Bin Cui

    ICDE, 2020

  • A Reinforcement Learning-based Method for Join Optimization

    Xinyi Zhang, Zhipeng Zhang, Bin Cui

    NDBC, 2020 (Best Student Paper)

  • PS2: Parameter Server on Spark

    Zhipeng Zhang, Bin Cui, Yingxia Shao, Lele Yu, Jiawei Jiang, Xupeng Miao

    SIGMOD, 2019

  • MLlib*: Fast Training of GLMs using Spark MLlib

    Zhipeng Zhang, Jiawei Jiang, Wentao Wu, Ce Zhang, Lele Yu, Bin Cui

    ICDE, 2019

  • An Experimental Evaluation of SimRank-based Similarity Search Algorithms

    Zhipeng Zhang, Yingxia Shao, Bin Cui, Ce Zhang

    VLDB, 2017

Awards

Blogs

Selected public posts will be listed here.