Designing an Educational Virtual Assistant Based on Agentic AI for Student Diagnosis and Individualized Remediation

The Advent of Generative AI and most recently AI Agents has significantly impacted all major fields especially education, bringing many advantages for both students and educators. Nevertheless, student diagnosis, individual remediation and teacher-centric solutions are still overlooked by researchers in AIED. Even schools often neglect the importance of cognitive diagnosis and student monitoring to the student learning process, let alone integrating them in a real classroom setting using these emerging technologies. In this paper, we study the recent work regarding educational virtual assistants and student diagnosis, while identifying some of the challenges. Thus we propose a new framework made for student diagnosis and remediation powered by an AI virtual teacher assistant based on Vision Language Models (VLM), Large Language Models (LLM) and Multi-Agent Systems (MAS).