Yu SUN

Yu SUN (孙宇)

Machine Learning Engineer @ Weixin Group, Tencent

universuen@gmail.com sun.yu@u.nus.edu

About Me

I am a Machine Learning Engineer at Weixin Group, Tencent. My current work focuses on the full lifecycle of Multimodal Large Language Models (MLLMs)—from architecting pre-training strategies to fine-tuning and post-training optimization—to enhance content understanding and safety within the WeChat ecosystem.

I hold a Master's degree in Artificial Intelligence from the National University of Singapore (NUS) and a Bachelor's degree in Cyberspace Security from Wuhan University (WHU). Previously, I was a Machine Learning Engineer Intern at ByteDance Singapore, working on TikTok content safety.

My research interests include MLLMs, Generative Models, Content Understanding, and Out-of-Distribution Generalization.

Technical Expertise

MLLMs Model Training Framework SFT RLHF Content Safety & Risk Detection Generative Adversarial Networks (GANs) Diffusion Models PyTorch / DeepSpeed

Experience

Weixin Group, Tencent 2024-Present
Machine Learning Engineer
Focusing on WeChat content safety by developing and deploying Multimodal Large Language Models (MLLMs) for advanced content understanding and risk detection.
TikTok, ByteDance Singapore 2022-2024
Machine Learning Engineer Intern
Worked on machine learning solutions for TikTok content safety and moderation.

Education

National University of Singapore (NUS) 2022 - 2024
M.Comp. in Artificial Intelligence
Wuhan University (WHU) 2018 - 2022
B.Eng. in Cyberspace Security

Publications

* Indicates Equal Contribution / Co-first Authorship

Neural Comput & App
VGAN-BL: imbalanced data classification based on generative adversarial network and biased loss
H Ding*, Y Sun*, N Huang, X Cui
Neural Computing and Applications (2024)
IEEE TIFS
TMG-GAN: Generative adversarial networks-based imbalanced learning for network intrusion detection
H Ding, Y Sun, N Huang, Z Shen, X Cui
IEEE Transactions on Information Forensics and Security (2023)
Info Sciences
RVGAN-TL: A generative adversarial networks and transfer learning-based hybrid approach
H Ding*, Y Sun*, N Huang, Z Shen, Z Wang, A Iftekhar, X Cui
Information Sciences (2023)
IP&M
RGAN-EL: A GAN and ensemble learning-based hybrid approach for imbalanced data classification
H Ding*, Y Sun*, Z Wang, N Huang, Z Shen, X Cui
Information Processing & Management (2023)
Preprint
Towards equivariant graph contrastive learning via cross-graph augmentation
Z Liu, A Zhang, Y Sun, Y Li, Y Shi, S Li, X Wang, X He, TS Chua
arXiv preprint (2023)

Featured Projects

hurricore
A universal deep learning training framework built on top of PyTorch and Hugging Face Accelerate.