CV

Education

• M.S. in Comuter Science and Technology, Shanghai Jiao Tong University, 2022 - 2025
• B.Eng. in Transportation Engineering, Tongji University, 2018 - 2022

Work experience

• 2023.8 - Present: Research & Engineering Intern
  • ByteDance Infrastructure Department

  • Duties included:

    1) Contributed to the development of workload series forecasting algorithm for Horizontal Pod Autoscaling in Kubernetes (K8s). Designed a period detection algorithm to assist decomposition-integration forecasting strategy. This algorithm has been deployed in Volcano Kubernetes Engine, resulting in increased revenue of the million level.

    2) Led the development of a comprehensive and user-friendly universal time series platform. This platform offers users a convenient interface to access time series related algorithms. Took charge of the entire stack design, including integrating deep models into the algorithm side, implementing the server side using Flask, and building the front-end with React.

    3) Designed the deep forecasting model ATFNet, which adaptively utilizes the time-frequency domain Transformer for predicting various periodic dynamic time series. It has achieved leading performance on both internal workload data and public datasets. The preprint version’s repository has already garnered 100+ stars on GitHub. Link: https://github.com/YHYHYHYHYHY/ATFNet.

    4) Led the development of a forecasting algorithm for predicting the submission volume of TikTok videos and photos. The algorithm has been successfully implemented in a real - world application. Through evaluation in the first three quarters of 2024, it has achieved an average forecasting accuracy exceeding 90% across all regions.

    5) Led the design of an advanced optimization scheme for Kubernetes HPA’s proactive scaling, integrating online frequency domain prediction methods with dynamic queue modeling. This end-to-end (prediction to decision) scaling plan recommendation aims to achieve automatic scaling decisions with reduced service latency and resource overhead.

• 2023.2 - 2023.7: Research Intern
  • Huawei Cloud Computing Technology

  • Duties includes:

    1) Led the design of the CloudChurn model for identifying enterprise user churn in cloud service scenarios, completing performance tests that exceeded acceptance criteria by 33.85%.

    2) The CloudChurn model has been deployed on Huawei Cloud’s ModelArts platform and, after four months of online operation, achieved an average user churn identification accuracy of 78.3%.

Skills

• Programming Language: Python, C/C++
• System Skills: Linux, Kubernetes, Docker
• Framework: PyTorch
• Language Proficiency: IELTS 7.0

Personal Interest

• I am actively involved in soccer and typically play in the positions of CM and Winger.

• I’ve been an amateur guitarist for nine years now. My musical preferences predominantly lie in Blues, Prog-Rock, and Jazz Fusion. Some of my favorite bands are Muse, Radiohead, and T-Square.

• I am an experienced League of Legends player with a personal best rank of Diamond II. I mainly specialize in the Jungle and Support.