<<<<<<< HEAD
=======

Lei Zhao

📍San Marcos, TX, USA

[Github] [LinkedIn] [x.com] [Email] [Website]

>>>>>>> e8cb41537dec4c6d23a129f6adf87d72c9ec763b

Lei Zhao 赵雷

I am a Computer Science PhD student at Texas State University, advised by Dr. Chul-Ho Lee [Website]. My current research focuses on Machine Learning Systems (MLSys) and Graph Neural Networks (GNNs).

I have a bit of a hacker past that defines my personality. I prefer building infrastructure from scratch to master every detail, and I like to squeeze every bit of performance out of hardware.

Lei Zhao

Education

Aug 2024 - Present

Texas State University

Ph.D. Student in Computer Science (Advisor: Chul-Ho Lee)

Machine Learning Systems, Graph Neural Networks

Sep 2021 - Jun 2024

China University of Geosciences

M.Eng. in Electronic and Information (Advisor: Longsheng Wei)

Computer Vision, Image Quality Assessment

Sep 2008 - Jun 2012

Hubei University of Technology

B.Eng. in Electrical Engineering and Automation

Publications

2025

A Pavement Crack Segmentation Method Based on Deformable Convolution and Enhanced Perceive Network. [Paper]

Lei Zhao, Longsheng Wei, Zhi Ma, Zhiheng Liu. Multimedia Tools and Applications, 84(19):20995-21016.

  • Designed the network structure based on DCN and customized UperNet.
  • Reduced the UperNet layers and simplified the PPM head for local texture details.
  • Adopted deformable convolution for strip-like pavement cracks.
  • Fused features from different heads with trainable channel weights.
2023

Single Human Parsing Based on Visual Attention and Feature Enhancement. [Paper]

Zhi Ma, Lei Zhao, Longsheng Wei. Journal of Advanced Computational Intelligence and Intelligent Informatics, 27(4):561-566.

  • Designed the network structure based on Cross-Attention and ResNet-101.
  • Added an edge sensing module for better edge perception.
2021

Reduced Reference Quality Assessment for Image Retargeting by Earth Mover's Distance. [Paper]

Longsheng Wei, Lei Zhao, Jian Peng. Applied Sciences, 11(20):9776.

  • Designed the network structure based on SIFT and saliency features.

Projects

Nov 2022 - Apr 2023

Building Low-Cost High-Performance GPU Computing Servers for Laboratory

  • Purchased and assembled hardware including motherboard, PSU, RAM, GPU, PCIe splitter, SSD, and HDD.
  • Installed Xubuntu and configured CUDA, Docker, proxy rules, NAT traversal, and X11 forwarding.
  • Set up a dataset server and shared directory with NFS.
  • Created Dockerfiles for frequently used environments.
  • Benchmarked hardware combinations across CPU families, PCIe lanes, PCIe generations, and NVIDIA GPUs.
  • Tuned performance with htop, nvtop, iotop, top, bcache, and swap.
GPU computing cluster Server motherboard
Jun 2022 - Aug 2022

EEG and EOG based Real-Time Driver Fatigue Detection [GitHub]

China Graduate Electronics Design Contest

  • Developed a module to translate MATLAB data to Python pickle dumps.
  • Designed a CNN-based network to fuse EEG and EOG features for real-time driver fatigue detection.
  • Prepared slides and poster for demonstration.
Apr 2022 - Jun 2023

A Malfunction Detection System for Guangzhou Metro System

  • Developed image preprocessing to combine original images and generate target-area images.
  • Adopted YOLOv5 and Fast R-CNN in the malfunction detection process.
  • Developed a module to fuse recognition results and provide a Flask API.
  • Composed Docker Compose files for instant deployment across frontend, backend, and database modules.

Experience

Jun 2016 - May 2020

Wuhan Stack Tech Co., Ltd.

Founder, Developer

  • Developed websites, Chrome extensions, and WeChat applets.
Jul 2012 - May 2016

Ericsson

Radio Access Network Engineer

  • Worked on system integration, network optimization, and signal analysis.
  • Verified features against 3GPP specifications.
Sep 2010 - Jun 2011

HGDonline.net Coding Club

Web Developer

  • Participated in the revision of the university BBS.
  • Developed 59score by reverse-engineering the undergraduate registrar system, enabling students to check grades.

Skills

  • Deep Learning Systems: Pipeline bubble optimization, PyTorch profiling, Distributed Data Parallel (DDP), DeepSpeed, 3D parallelism.
  • Programming & Frameworks: Python (PyTorch, OpenCV, TraceViz), C/C++ (OpenMP), CUDA, shell scripting.
  • Systems & Infrastructure: Linux, Docker Compose, Git, network analysis (Wireshark), Slurm.
  • Web & Tools: JavaScript (AngularJS, Vue.js), PHP (Laravel), API testing (Postman), HTTP debugging (Fiddler), MySQL.

Awards

  • China University of Geosciences: First Class Scholarship twice, Second Class Scholarship once, and Third Class Prize in the China Graduate Electronics Design Contest.
  • Hubei University of Technology: Second Class Scholarship once, Third Class Scholarship twice, and Outstanding Student.

Patents

  • A Human Body Posture Estimation Method and System Based on Shape Similarity. CN202210752411.0
  • A Reference-Free Image Quality Evaluation Method Based on Graph Convolution and Multi-scale Features. CN202310801526.9
  • A Twin Network Single Target Tracking Method and Device. CN202310897743.2
  • Real-time Semantic Segmentation Method, Device, Electronic Equipment, and Storage Medium for Night Images. CN202310856554.0