About Me
Hello! I am Li Xiang, a PhD student at Nanjing University. I received my Master’s degree from the University of Electronic Science and Technology of China, where I was supervised by Prof. Lixin Duan and Prof. Yanwu Xu.
My research focuses on post-training adaptation of medical AI models, spanning lightweight deep learning networks, large-scale foundation models, and inference acceleration for large models. I work on domain adaptation, test-time training, and reinforcement learning to improve the reasoning-time robustness, efficiency, and generalizability of visual, language, and multimodal systems in real-world medical scenarios.
Education
Nanjing University (NJU)
Ph.D. Student
Present
University of Electronic Science and Technology of China (UESTC)
Master’s Degree
Graduated
Experience
OneFlow
Engineer Intern (Deep Learning Framework R&D)
2021 - 2023
SiliconFlow
Research Intern (Diffusion Model Inference Acceleration)
2023 - 2025
Projects
:
An out-of-the-box acceleration library for diffusion models, providing fast integration with tools such as HF diffusers and ComfyUI, along with optimized GPU kernels and PyTorch compilation support for production-ready inference acceleration.
:
A user-friendly, scalable, and efficient deep learning framework featuring a PyTorch-like API, n-dimensional parallel training via Global Tensor, and graph compiler support for acceleration and deployment.
Awards
University of Electronic Science and Technology of China