Yinhan Zhang

🎓 PhD Student at HKUST

Portrait of Yinhan Zhang

Hi! I am Yinhan Zhang, a PhD student at HKUST. My work focuses on world models, real-time interactive video generation, and image and video generation.

I enjoy connecting research ideas with real-world systems, from generative-model training and data pipelines to multimodal products.

Research interests

  • World Models & Real-Time Interactive Video Generation
  • Image and Video Generation

News

  • Jun. 2026One paper was accepted to ECCV 2026.
  • Mar. 2026One paper was accepted to ACM SIGGRAPH 2026.
  • Feb. 2026One paper was accepted to CVPR 2026.
  • Jan. 2026One paper was accepted to ICLR 2026.
  • Jun. 2025One paper was accepted to ICCV 2025.

Education

  • Aug. 2026
    — Present    HKUST

    The Hong Kong University of Science and Technology

    PhD · Artificial Intelligence

  • Sep. 2024
    — Jul. 2026    HKUST

    The Hong Kong University of Science and Technology

    MPhil · Artificial Intelligence

  • Sep. 2020
    — Jun. 2024    Taiyuan University of Technology

    Taiyuan University of Technology

    B.Eng. · Data Science and Big Data Technology

Experience

  • Jan. 2026
    — Present    Tencent

    Tencent Games

    Research Intern

    Research on Game World Models.

  • Nov. 2024
    — Nov. 2025    China Telecom

    TeleAI · China Telecom AI Research Institute

    Video Generation Research Intern

    Controllable human-motion generation, VAST 1.0 foundation-model development, and end-to-end video data pipelines.

  • May 2024
    — Jul. 2024    Tencent

    Tencent Hunyuan

    Research Intern

    Text-image alignment, query rewriting, multi-round dialogue image generation, and model-routing strategies.

  • Nov. 2023
    — Apr. 2024    JD.com

    JD Retail

    MLLM Algorithm Intern

    Built text and multimodal product-matching systems with hard-negative mining and attribute-aware alignment.

  • Jul. 2023
    — Nov. 2023    Baidu

    Baidu Financial Services

    NLP Algorithm Intern

    Developed intent recognition and an intelligent customer-service model for financial-domain conversations.

Selected Projects

Pyoh Technology programming education hardware

Pyoh Technology · Programming Education

Interactive hardware that helps children learn programming through play, combining creativity education with hands-on computing.

Project ↗
Tencent Open Source contribution project

Tencent Rhinoceros Bird Open Source Program

Implemented Struct2vec on Tencent Angel and evaluated it against DeepWalk and Node2vec for graph clustering and classification.

Immersive Liangzhu AR project

Immersive Liangzhu · AR Interactive Series

An AR museum experience combining generative restoration of ancient characters with a Liangzhu knowledge-base dialogue system.

Honors & Awards

  • 2024Technical Application Prize, Alibaba Cloud Large Model Agent Challenge · Top 20.
  • 2023 / 24National Third Prize, China Mobile Wutong Cup · Top 10 · Team Leader.
  • 2023National Third Prize, Chinese Robotics and Artificial Intelligence Competition · Team Leader.
  • 2022National Second Prize, Internet+ Innovation and Entrepreneurship Competition.
  • 2022National Second Prize, China Software Cup.
  • 2022Outstanding Student, Tencent Rhinoceros Bird Open Source Program.

Open Source

  • LLM Cookbook ↗★ 24.8kA practical large-language-model handbook for developers.
  • D2L Solutions ↗★ 609Community solutions and learning notes for Dive into Deep Learning.
  • Grape Book ↗★ 287An open Chinese learning resource for graph deep learning.