As part of a 10-week Independent Study course, Reinforcement Learning was studied and used to control a simple ball-balancing system. A Simulink Model of the system was developed using mathematical modeling and System Identification techniques. The implementation of Reinforcement Learning in MATLAB was studied and many different learning strategies were experimented with. A DDPG Agent was found to converge on a control strategy during training. A Reinforcement Learning Agent was deployed on the hardware but produced unsatisfactory results. Performance comparisons were made between the Reinforcement Learning Agent and a Proportional-Derivative controller. MATLAB and Simulink were used throughout the project.