Shuo Cai has actively engaged in research projects that tackle real-world challenges through machine learning applications. While at South China University of Technology, he developed a hybrid path planning framework for underwater robots by integrating optimization algorithms with deep reinforcement learning, significantly enhancing navigation efficiency in complex marine environments. Currently, as a technical staff intern under Professor Hongxia Yang, Shuo is focused on advancing large language models (LLMs). His work involves improving their reasoning capabilities in tasks such as coding and multimodal problem-solving. His efforts include fine-tuning models using supervised fine-tuning and reinforced learning to align task-specific objectives, designing efficient inference pipelines for real-time deployment, and exploring lightweight frameworks to adapt LLMs to various scenarios while maintaining interpretability and scalability.