Tsinghua University · Ph.D. Student
Electronic and Information Engineering, Institute of Data and Information, Tsinghua Shenzhen International Graduate School, Tsinghua University.
Ph.D. student in Electronic and Information Engineering at the Institute of Data and Information, Tsinghua Shenzhen International Graduate School, Tsinghua University, advised by Prof. Zhiyong Wu . I am a member of the THUHCSI Lab. My research interests include Speech Large Language Models, Audio Codec, Audio-Driven 3D Facial Animation Generation, and Stereo Image Compression.
Electronic and Information Engineering, Institute of Data and Information, Tsinghua Shenzhen International Graduate School, Tsinghua University.
Dual-degree undergraduate with a bachelor's degree in Software Engineering and a minor in Business Administration. Ranked 1st in the major and received the National Scholarship twice.
Co-first author. Proposed a frequency-domain alignment module and a bidirectional multi-reference entropy model for efficient stereo image compression.
Speech Algorithm Engineer. Responsible for foundation-model training for speech large language models and participated in the training and release of VoxCPM .
Responsible for model training and the end-to-end evaluation and optimization pipeline for audio-driven 3D facial animation. As a core contributor, participated in the release of Digital Jingdezhen · Porcelain Craftsman , contributing to Jingdezhen's successful World Heritage inscription.
Conducted research on neural audio codecs, audio discretization, and representation learning, and developed the DynamicCodec experimental platform.
Conducted research on stereo image compression and developed FreqSIC , published at CVPR 2026 as a co-first author and selected as a Highlight paper.