AI and LLM Systems

Built and deployed a course-specific conversational AI tutor for university statistics, comparing Socratic and instructional tutoring strategies in a field experiment and deriving design guidelines for controllable LLM-based coaching systems.

Developed and evaluated an AI chatbot that personalizes standardized digital competence assessments based on individual user profiles, revealing that adaptive question generation improves relevance for advanced users while beginners prefer standardized formats.

Developed and evaluated a human-AI collaboration system that uses adaptive text feature recommendations to improve the quality of peer feedback in large-scale university courses, tested with 490 students across three institutions.