About Me

Welcome to my academic page! I am an undergraduate student currently pursuing a B.S.E in Computer Science at the University of Michigan, Ann Arbor, and a B.Eng. in Mechanical Engineering at Shanghai Jiao Tong University.

My research interests lie in the intersection of Robot Learning (VLA, VAM), Agentic Systems, and Efficient Machine Learning. My goal is to develop highly capable and intelligent systems that can reliably perceive, reason, and act in complex environments.

Research Experience

  • Efficient Inference for Embodied Foundation Models (Nov 2025 – Present)
    Research Assistant, Advisor: Dr. Jiachen Liu (University of Michigan)
    Investigating efficiency improvements in embodied foundational models. My work involves profiling COSMOS-Policy, validating cross-attention KV caching strategies across dozens of RoboCasa tasks, and designing task-aware token compression to selectively prune image tokens without sacrificing task success rates.

  • Gravitational Effect on Swarming Behavior of Microorganisms (Sep 2024 – Aug 2025)
    Research Assistant, Advisor: Prof. Zijie Qu (Shanghai Jiao Tong University)
    Trained learning-based models for colony boundary segmentation, diagnosing failure modes in initial U-Net models and ultimately adopting a robust SAM-based pipeline for reliable automated detection.

Selected Projects

  • Web Agent Version-Robust Benchmark (Jan 2026 – Present)
    Building a reproducible benchmark pipeline using Docker to test LLM-based web agents (e.g., QWen3-VL-30B) across historical snapshots of open-source websites. Formulating taxonomies for version-induced failure modes and implementing Knowledge Graph-based methods for automated task generation.
  • Probabilistic Motion Planning for Redundant Robots (Nov 2025 – Dec 2025)
    Reproduced and benchmarked sampling-based motion planning algorithms (RRT-Connect, PRM) for a 7-DOF Franka Panda in simulation. Developed a hybrid sampling strategy that reduced trajectory generation latency by ~75%.

Technical Skills

  • Programming Languages: Python, C/C++, MATLAB, Java
  • Tools & Frameworks: PyTorch, Hugging Face, Linux, Docker, Git, LaTeX, SolidWorks