Lunjie Zhu 朱伦杰

Ph.D. Student at HKUST

Hong Kong SAR

Email Scholar GitHub

I am a Ph.D. student in Electronic and Computer Engineering at the Hong Kong University of Science and Technology (HKUST), advised by Prof. Jun Zhang. I received my B.Eng. in Information Science and Technology from Harbin Institute of Technology, Shenzhen.

My research focuses on building efficient and high-performing multimodal generative models. Currently, I explore scalable and high-fidelity pre-training architecture design for video generation foundation models, including pixel-space diffusion. My prior work spans long-horizon, real-time streaming joint audio-video generation, video VAE decoder acceleration, interactive world model and so on.

I am always happy to chat about research and potential collaborations — feel free to reach out.

Aside from work and studying, I am a sports lover, including, but not limited to, badminton 🏸, running 🏃, and cycling 🚴.

News

2026.10: Join ByteDance as a research intern.
2026.10: Our paper RIGen is released.
2026.09: Our papers Salt++ and Radia are released.
2026.08: Our paper Omni-LiveAvatar is released.
2026.05: 🎉 Our Flash-VAED is accepted by ICML 2026.
2026.04: Join Vivix AI as research intern.
2026.03: Our papers VLMQ and PrismMirror are released.
2026.02: Our paper Flash-VAED is released.
2025.09: Join iComAI Lab as Ph.D. student.

Publications | Full List

*/Conference/*

	Flash-VAED: Plug-and-Play VAE Decoders for Efficient Video Generation Lunjie Zhu, Yushi Huang, Xingtong Ge, Yufei Xue, Zhening Liu, Yumeng Zhang, Zehong Lin, Jun Zhang International Conference on Machine Learning (ICML), 2026 [paper] [code] [models]

*/Preprints/*
	Omni-LiveAvatar: Minute-Level Real-Time Streaming Joint Audio-Visual Avatar Generation Lunjie Zhu, Xingtong Ge, Fangyu Lin, Yi Zhang, Zhening Liu, Mengfei Li, Yumeng Zhang, Guanglu Song, Yu Liu, Jun Zhang [paper] [code]

	RIGen: Beyond Action Following with Reciprocal Interaction in Generative Gameplay Mengfei Li, Yi Zhang, Xingtong Ge, Lunjie Zhu, Fangyu Lin, Peng Li, Guanglu Song, Yu Liu, Wei Xue, Wenhan Luo, Yike Guo [paper] [project] [code] [models]

	Salt++: Context-Aligned Post-Training for Few-Step Streaming Multimodal Generation Xingtong Ge, Yutong Wang, Lunjie Zhu, Haitao Lin, Fangyu Lin, Yushi Huang, Xin Zhang, Yi Zhang, Yu Liu, Jun Zhang [paper] [code]

	Enhancing Autoregressive Video Generation via Representation Adversarial Distillation Fangyu Lin, Xingtong Ge, Lunjie Zhu, Yi Zhang, Zhening Liu, Tianhang Wang, Mengfei Li, Yumeng Zhang, Guanglu Song, Yu Liu, Jun Zhang [paper]

	Real-Time Human Frontal View Synthesis from a Single Image Fangyu Lin, Yingdong Hu, Lunjie Zhu, Zhening Liu, Yushi Huang, Zehong Lin, Jun Zhang [paper] [code]

	VLMQ: Efficient Post-Training Quantization for Large Vision-Language Models via Hessian Augmentation Yufei Xue, Yushi Huang, Lunjie Zhu, Jiawei Shao, Jun Zhang [paper]