Experience

**Assistant Professor**

Shenzhen University

Building a research group focused on VLM edge-device inference acceleration and VLM interpretability. The group is actively recruiting master’s students and interns year-round.

AI Researcher

HONOR Device Co., Ltd.

  • Led AI innovation for HONOR’s gaming business. Built M-DLSS 3.0, the industry’s first NPU-GPU heterogeneous frame interpolation system shipped to consumer phones.
  • Led the embodied intelligence capabilities of RobotPhone.
  • Close research and engineering collaboration with Google Gemma, Qualcomm, and Alibaba Qwen Lab.

Visiting Researcher

National University of Singapore

Research exchange on vision-language understanding.

Visiting Researcher

Tsinghua University

Research exchange on 3D human motion understanding.

Education

PhD in Computer Applied Technology

University of Chinese Academy of Sciences (UCAS)

Research on 3D human pose estimation, motion capture, and motion generation. Published at ICCV, ECCV, ACM MM, IEEE TCSVT, and other top venues.

Visiting Researcher

National University of Singapore (NUS)

Research exchange on vision-language understanding.

Visiting Researcher

Tsinghua University

Research exchange on 3D human motion understanding.
Skills & Hobbies
Research
Vision-Language Models
Edge / On-device Inference
Model Interpretability
Embodied AI
Generative Models (Diffusion, VAE, SSM)
Engineering
PyTorch
NPU-GPU Heterogeneous Computing
Real-time Inference Optimization
Mobile / Edge Deployment
Awards
Industry Collaborations
Google · Qualcomm · Alibaba ∙ Present
Active research collaborations with Google Gemma, Qualcomm, and Alibaba Qwen Lab during HONOR tenure.
Shipped to consumers: M-DLSS 3.0
HONOR Device ∙ Present
Industry-first NPU-GPU heterogeneous frame interpolation system shipped on HONOR gaming phones.
Reviewer
ICLR · ICCV · ECCV · ACM MM ∙ Present
Serving as a reviewer for top-tier AI/CV conferences.
20+ Publications · 10+ Patents
ICCV / ECCV / ACM MM / IEEE TCSVT / IEEE TMM ∙ Present
Published at international top conferences and journals.
Languages
100%
Chinese Native
80%
English Fluent