PhD student Tianxin Wei has been awarded an Amazon Post Internship Fellowship, which will provide $20,000 in unrestricted funds and $20,000 in Amazon Web Services (AWS) credits to support Wei's research with his advisor, Professor Jingrui He. For the past two summers, Wei has served as an applied scientist intern at Amazon in Palo Alto, California. He has been part of a team that is working on search query understanding within Amazon apps and services, as well as developing shopping foundation models.
Wei's research at Illinois primarily focuses on trustworthy machine learning and transfer learning, especially in the context of knowledge transfer across different domains and modalities. According to Wei, his work at Amazon over the past two years has been "highly relevant" to his doctoral research.
"In my first internship, I worked on multimodal learning—Amazon products contain multiple modalities of information, such as various text formats, images, and user interaction histories. My work focused on effectively integrating these sources of information and utilizing large models to provide users with personalized services like search and recommendation," he said. "In my second internship, I focused on multimodal safety, particularly exploring methods to prevent AI-generated images from causing harm in search and shopping experiences."
Wei continued his work with Amazon into the fall semester, in both of his internships. He earned his BE in computer science from the University of Science and Technology of China.