Yuxuan Xia

Master’s student at UC Santa Barbara (ECE), focusing on Vision Language Models and scalable AI systems.

Hello, I’m Yuxuan Xia.

I am currently a Master’s student in the Department of Electrical & Computer Engineering at the University of California, Santa Barbara. I completed my Bachelor of Science in Computer Science through a joint program between Southwest Jiaotong University and the University of Leeds, graduating with First Class Honours from the University of Leeds.

Currently, I am a Student Researcher at Peng Li’s Lab, where I focus on investigating and mitigating hallucinations in Vision Language Models (VLMs) by correcting erroneous attention mechanisms. I also work as an AI Data Annotator (CS Fellow) at Handshake AI, constructing high-level computer science datasets for LLM instruction tuning.

My research and project experience spans both machine learning and software engineering:

  • Machine Learning & AI: I have designed and implemented AI agents for healthcare using LangChain and RAG, implemented Zero-Order (ZO) optimization for gradient-agnostic models, and researched optimization techniques for Tensorized Transformers.
  • Algorithm & Software Development: I have worked as an Algorithm Engineer designing TDOA (Time Difference of Arrival) localization algorithms using MATLAB and C++. I also have full-stack experience developing with Vue.js and Spring Boot, and experience deploying YOLOv5s models on edge devices like the Nvidia Jetson Nano.

I am proficient in programming languages including Python, C/C++, Java, and JavaScript, and skilled in frameworks such as PyTorch, CUDA, Hugging Face, Vue.js, and Spring Boot.