This project addresses important challenges in China’s foreign affairs and business sectors, such as the shortage of skilled small-language speakers and limited teaching resources, as well as the difficulty of creating effective, real-life learning scenarios. It focuses on key technologies for crowdsourcing the collection of multilingual and multimodal language data, and for enabling smart interactions with virtual teachers. Specifically, the project aims to build a standardized system for collecting this multilingual data through crowdsourcing; create detailed and realistic virtual teacher avatars by combining human shape and appearance features; and develop AI-powered interaction technologies that let these virtual teachers adjust dynamically to each learner’s needs. By solving these core technical problems, the project supports smarter language education for diplomatic and business purposes and helps train more skilled small-language professionals across China.