Enhancing an AI-Powered Study Assistant: Design and Evaluation of New Capabilities for Personalized and Effective Learning

This thesis presents the design, implementation, and evaluation of new capabilities for an AI-powered conversational tutor embedded in the FACEIT e-learning platform. Building on the existing architecture, the work explores several directions for improving the assistant's overall effectiveness, including how it retrieves and makes use of relevant information to support students and teachers, alongside broader extensions to its capabilities within the platform. The proposed enhancements are integrated into the existing system and evaluated through a combination of technical assessment and user-facing evaluation, considering both feasibility and impact on the overall learning experience. The thesis aims to demonstrate how a set of targeted improvements can make a conversational tutor more effective and useful within a real e-learning environment.