Aplikace "stylu Duolingo" pro přípravu na zkoušky z matematiky pro ekonomy

Mathematics for economists is consistently one of the highest-failure-rate courses at the Prague University of Economics and Business, and students lack a tool that would enable systematic practice in short intervals with immediate feedback. This thesis designs and implements a web application, Mathingo, which transposes the proven gamification principles of Duolingo into the context of exam preparation for the 4MM101 course. Application requirements were formulated through mixed-methods research combining semi-structured interviews with mathematics department faculty and a quantitative student survey. The application was developed using the vibe coding methodology with the Claude Code tool, whose selection was justified through an analysis of the current ecosystem of AI-assisted development tools. The resulting application is deployed on a production domain and implements all eight formulated functional requirements. The thesis also reflects on the process of working with an AI agent across 21 recorded work sessions and identifies seven classes of recurring problems. A specific contribution of this work is the systematic anti-dark-pattern design of the notification system, which represents a conscious opposition to the manipulative practices of the reference product and thereby constitutes a critical revision of the originally formulated functional requirement.