[2023.9] I'm joining Bytedance as a research scientist.

I'm a Research Scientist at ByteDance, working on monitoring and troubleshooting large-scale RDMA and GPU datacenter networks, especially for LLM. My research interests are widely on distributed systems, network systems, cloud systems, and datacenter management. I've been working on the specific areas like:

• Improve observability and reliability for large scale systems, through distributed tracing and root cause analysis.
• Quantify, understand, and analyze performance for data placement in distributed cache/memory systems.
• Develop other large-scale applications e.g. video streaming and secure storage systems.
• RDMA and GPU datacenter network management and troubleshooting for LLM.

My research has been awarded with an ACM SIGMETRICS Kenneth C. Sevcik Outstanding Student Paper Award. I was a postdoc researcher at Princeton University, working with Prof. Ravi Netravali. Before that, I received my Ph.D. from Emory University in 2021 (working with Prof. Ymir Vigfusson), Master from Georgia Tech in 2017(was in Ph.D. program, worked with Prof. Karsten Schwan memorial page ), and Bachelor from Tsinghua University in 2015. I transferred to Emory in 2018 as a post-qualified Ph.D. student.

Thesis:

Papers:

Yazhuo Zhang, Rebecca Isaacs, Yao Yue, Juncheng Yang, Lei Zhang, Ymir Vigfusson
In ACM SoCC 2023
  Paper   The Benefit of Hindsight: Tracing Edge-Cases in Distributed Systems

Lei Zhang, Vaastav Anand, Zhiqiang Xie, Ymir Vigfusson, Jonathan Mace
In USENIX NSDI 2023
  Paper     Slides     Code     Benchmark Code

Lei Zhang, Juncheng Yang, Anna Blasiak, Mike McCall, Ymir Vigfusson
In USENIX HotCloud 2020
  Paper     Slides     Video

Lei Zhang, Reza Karimi, Irfan Ahmad, Ymir Vigfusson
In ACM SIGMETRICS 2020
  Paper     Video
  Kenneth C. Sevcik Outstanding Student Paper Award

Lei Zhang, Douglas Blough
In IEEE International Conference on Dependable Systems and Networks (DSN) 2018
  Paper

Maomeng Su, Lei Zhang, Yongwei Wu, Kang Chen, Keqin Li
In IEEE Transactions on Computers 2016, 65(6): 1964-1977.
  Paper

In preparation: