CAREER: Orchestrating Generative AI-based Agents for Personalized and Scalable Computing Education

Computing education is growing rapidly, yet instructors often struggle to provide timely and personalized feedback to large numbers of students. Generative artificial intelligence (AI) has the potential to support learning, but students may rely on it in ways that limit deep understanding, while instructors often lack visibility into how it is used. This project investigates how to combine AI capabilities with human guidance to support effective and responsible learning in computing classroom settings. The goal is to help students receive meaningful feedback while enabling instructors to monitor and guide AI-supported learning. By improving how feedback is delivered and supervised, the project aims to support student learning, promote responsible use of AI, and make high-quality computing education more personalized and scalable. The results will contribute to computing education and help prepare an AI-ready workforce. This project develops a novel multi-agent framework that coordinates three types of intelligent components to facilitate AI-supported learning. Teaching assistant agents powered by large language models provide personalized, pedagogically aligned, and multimodal feedback, guided by principles that determine what feedback to provide, when to provide it, and how to present it. An instructor-facing analytics system and an instructor agent are used to model and analyze student–AI interactions, identifying patterns such as disengagement, misuse, or ineffective feedback. Finally, an orchestration agent, together with real-time interfaces, allows instructors to monitor classroom activity and dynamically adjust the behavior of the other AI-based agents during instruction. The project draws on methods from human-computer interaction, machine learning, and the learning sciences, including multimodal interaction, visual analytics, large language model-based feedback generation, and human-in-the-loop system design. The research will be evaluated through classroom deployments, mixed-method studies, and model performance analysis. Project outcomes will be disseminated through publications and shared resources to support adoption in various educational settings. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria. NSF Award ID: 2540315 | Program: 01003031DB NSF RESEARCH & RELATED ACTIVIT,01002930DB NSF RESEARCH & RELATED ACTIVIT,01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Meng Xia | Institution: Texas A&M Engineering Experiment Station, COLLEGE STATION, TX | Award Amount: $383,884 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2540315 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2540315.html