The rapid evolution of digital learning platforms has transformed educational methodologies, with an increasing emphasis on personalized learning experiences. This thesis aims to personalize exercise feedback for students, leveraging their individual competencies and backgrounds by creating learner profiles. Building on existing (semi-)automated feedback mechanisms in Athena, this approach addresses the diverse and large student body at TUM, aiming to provide more instructive and personalized feedback. The project involves data collection, iterative development, and AI model integration to ensure efficient and scalable feedback personalization. By improving feedback quality and reducing lecturers’ workloads, this research seeks to enhance personalized education, helping students achieve their full potential.