Epidemic modeling, encompassing deterministic and stochastic approaches, is vital for understanding infectious diseases and informing public health strategies. This research adopts a prescriptive approach, focusing on reinforcement learning (RL) to develop strategies that balance minimizing infections with maximizing in-person interactions in educational settings. We introduce SafeCampus , a novel tool that simulates infection spread and facilitates the exploration of various RL algorithms in response to epidemic challenges. SafeCampus incorporates a custom RL environment, informed by stochastic epidemic models, to realistically represent university campus dynamics during epidemics. We evaluate Q-learning for a discretized state space which resulted in a policy matrix that not only guides occupancy decisions under varying epidemic conditions but also illustrates the inherent trade-off in epidemic management. This trade-off is characterized by the dilemma between stricter measures, which may effectively reduce infections but impose less educational benefit (more in-person interactions), and more lenient policies, which could lead to higher infection rates.