Developed and implemented a Reinforcement Learning agent utilizing the Deep Q-Network algorithm, PyTorch, and OpenAI Gym frameworks to play Atari Pong, showcasing AI and machine learning expertise. Conducted in-depth analysis to evaluate the effect of varying environmental factors, such as the size of the paddles, on learning process and agent performance, ensuring robustness and adaptability in different scenarios. Maintained constant hyperparameters, such as batch size, learning rate, and the epsilon start, decay, and minimum values, throughout the experimentation phase, enabling accurate comparisons and reliable conclusions. Observed consistent outperformance when trained on a wider variety of environments over a singular environment, and presented these findings to the School of Science, initiating knowledge-sharing within the academic community.