This project aims to develop a hands-on skill self-learning system based on artificial intelligence technology, using Basic Life Support (BLS) skills as the focus of system development. The system will leverage AI for motion recognition and real-time feedback, enabling users to complete self-directed learning without the need for continuous instructor supervision. A virtual rater function is integrated into the system to evaluate users' CPR performance through image recognition and motion analysis, providing detailed learning feedback and guidance. International students participating in internships at our laboratory will learn to design and validate the practical application of AI models, contribute to the development of educational platforms, and conduct reliability and validity testing of the system. This project not only addresses the global demand for scalable and accessible skill training solutions but also enhances students' understanding of AI applications in medical education while fostering cross-border knowledge exchange and collaboration. This system will serve as a prototype for self-learning applications in other operational skill training, driving the development of innovative competency-based education.