Efficient control of building energy systems is essential for achieving comfortable indoor environments and minimizing energy consumption. This study focuses on the development of an innovative reinforcement learning (RL) control agent, referred to as Inv_RL_Agent, specifically designed to optimize the performance of heating, ventilation, and air conditioning (HVAC) systems in residential buildings. Traditional control methods such as Proportional-Integral (PI) controllers, though widely used, often fail to balance energy efficiency and indoor comfort optimally. The Inv_RL_Agent addresses these limitations by learning control policies through trial and error, adapting dynamically to varying building conditions. To evaluate its performance, the agent was implemented in the Building Optimization Performance Test (BOPTEST) framework, a standardized simulation platform designed for assessing advanced control algorithms in building systems. Simulation results demonstrate that the Inv_RL_Agent outperforms traditional PI control strategies, delivering a 26% reduction in energy consumption while maintaining or improving indoor comfort levels based on established thermal comfort metrics. These findings underscore the potential of RL-based approaches to enhance energy management in residential buildings, offering a more adaptive and efficient alternative to conventional methods. The success of the Inv_RL_Agent highlights the promising role of artificial intelligence in the development of next-generation HVAC control systems that meet both sustainability and comfort objectives.