This study aims to develop a smart resume filter system using AI and ML to enhance the efficiency and accuracy of resume screening in talent acquisition. The methodology involves collecting job requirement texts from recruitment websites in the US and China, preprocessing data with NLP tools, and using Word2Vec and LDA for vocational qualification analysis and candidate matching. The system shows potential to improve candidate selection accuracy and efficiency. However, it underutilizes advanced AI models and focuses narrowly on textual data. Future research should integrate large-scale models and multimodal learning for a more comprehensive assessment. This study contributes by innovatively applying AI in recruitment, detailing the development methodology, and highlighting areas for improvement, including ethical considerations and continuous updates to feature extraction processes.