This thesis explores the implementation of a reinforcement learning agent within a 2D fighting game environment utilizing Unity and its machine learning toolkit. The primary objective is to conduct a comprehensive comparative analysis between the reinforcement learning agent and a traditionally scripted artificial intelligence system. The evaluation aims to discern the extent to which the integration of a reinforcement learning agent enhances the overall player experience within the gaming environment. Additionally, this research delves into the intricate process of creating the game itself, providing insights into the development challenges encountered throughout the project's lifecycle. By scrutinizing both the technical aspects of reinforcement learning in a gaming context and the practical hurdles faced during game creation, this thesis contributes valuable perspectives to the intersection of artificial intelligence, game design, and player engagement.