Shuai Yang
Student at Peking University
Beijing, China
Summary
🤩
Rockstar🎓
Top SchoolShuai Yang is a research-oriented machine learning engineer and PhD-trained researcher with nine years of experience focused on high-resolution portrait style transfer and generative models. Based in Beijing and affiliated with STRUCT@PKU and MMLab@NTU, he has hands-on contributions to influential projects such as DualStyleGAN and SIGGRAPH Asia’s VToonify, working on data preparation, face alignment, and training pipeline improvements. His background in applied mathematics informs a careful, quantitative approach to loss design and preprocessing for video and image stylization. Shuai blends academic rigor with practical engineering—modifying training scripts and tooling to make state-of-the-art style transfer methods reproducible and scalable. An underappreciated strength is his cross-domain experience aligning caricature/anime datasets and video face detection, which helps bridge research prototypes to real-world media applications.
9 years of coding experience
Master's degree, Applied Mathematics, A, Master's degree, Applied Mathematics, A at Peking University
Github Skills (15)
data-preprocessing10
face-detection10
computer-vision10
dlib10
stylegan10
machine-learning10
pytorch10
dataprep10
python10
image-processing10
style-transfer10
tensor8
modeling8
trainings8
tensorflow8
Programming languages (4)
Github contributions (5)
williamyang1991/VToonify
Sep 2022 - Jan 2023
[SIGGRAPH Asia 2022] VToonify: Controllable High-Resolution Portrait Video Style Transfer
Role in this project:
ML Engineer
ML EngineerContributions:165 commits, 2 PRs, 42 pushes in 4 months
Contributions summary:Shuai contributed to the VToonify project, which focuses on controllable high-resolution portrait video style transfer. Their commits primarily involve modifications to the training script, `train_vtoonify_d.py`, and related model files. These changes likely pertain to the training process, including loss functions, model parameters, and potentially data loading and preprocessing for style transfer models. They also updated related notebook files, inferring ongoing experimentation or modifications to the project.
portraitsiggraph-asia-2022siggraph-asiastyle-transferstylegan2
williamyang1991/DualStyleGAN
Mar 2022 - Jan 2023
[CVPR 2022] Pastiche Master: Exemplar-Based High-Resolution Portrait Style Transfer
Role in this project:ML Engineer
ML EngineerContributions:187 commits, 2 PRs, 61 pushes in 10 months
Contributions summary:Shuai primarily worked on data preparation and model training for the `dualstylegan` repository, which focuses on exemplar-based high-resolution portrait style transfer. Their contributions involved creating scripts for data alignment, specifically for caricature and anime datasets, and face detection in videos using dlib. They also modified code related to the style transfer process using the DualStyleGAN model.
pytorchportraitstyle-transferexemplarstylegan2
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.