zhouzj
Summary
🤩
RockstarZijun Zhou is a business analysis student at the University of Cincinnati with seven years of practical experience blending analytical tools like Excel and PowerPoint with hands-on engineering work in open-source ML projects. While studying business, Zijun has contributed as a back-end developer and ML engineer to PaddlePaddle/PaddleSlim, optimizing training configurations, improving memory efficiency, and enhancing model compression and pruning logic. That work shows a rare cross-disciplinary skill set: business analysis instincts paired with technical fluency in model optimization and latency prediction. Zijun is proactive about skill growth, continuously seeking opportunities to apply classroom learning to real-world codebases. Based in Cincinnati, they bring attention to resource-efficient solutions and a thoughtful approach to integrating new operators and fixing subtle sparsity bugs. Their GitHub motto, 明明如月，何时可掇, hints at a reflective mindset that balances technical rigor with cultural depth.
7 years of coding experience
Github Skills (12)
model-compression10
paddlepaddle10
pruning10
machine-learning10
python10
quantization9
computer-vision9
backend8
tensorrt8
data-structure5
algorithms5
data-structures5
Programming languages (2)
Github contributions (5)
PaddlePaddle/PaddleSlim
May 2021 - Jan 2023
PaddleSlim is an open-source library for deep model compression and architecture search.
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:70 reviews, 71 commits, 160 PRs in 1 year 8 months
Contributions summary:Zhouzj primarily focused on modifying and optimizing the codebase for deep model compression and architecture search. They adjusted batch sizes in multiple training scripts and configurations to address memory issues, demonstrating a concern for efficient resource utilization. The user also made significant contributions to the analysis module, adding op predictors for unknown operators and integrating new operations, which involved modifications to feature extraction and latency prediction functionalities. Furthermore, the user worked on fixing bugs related to model sparsity and pruning, particularly with respect to handling depthwise convolutions and removing unused nodes, indicating skills in model optimization and compression techniques.
berternieopenmmlabpaddleslimsparsity
zzjjay/PaddleSlim
Jun 2021 - Mar 2024
PaddleSlim is an open-source library for deep model compression and architecture search.
Contributions:6 PRs, 521 pushes, 212 branches in 2 years 9 months
pytorchdeep-learningcompressionopenmmlabpaddleslim
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.