This study demonstrates the performance improvement of an air-to-air 7.5 Ton (26.4 kW) rooftop unit (RTU) achieved by optimizing an evaporator's refrigerant circuitry using evolutionary algorithms. We measured its cooling capacity and Coefficient of Performance (COP) in an environmental chamber where we also measured the in-situ air velocity profile using Particle Image Velocimetry (PIV). We used a detailed heat exchanger model to simulate the performance of the original evaporator operating with the measured air distribution, and then we used an evolutionary algorithm-based optimization module embedded in the simulation tool to optimize the refrigerant circuitry for the measured inlet air distribution. The optimized refrigerant circuitry design was implemented in a new prototype evaporator, which replaced the original evaporator in the tested RTU system. Laboratory tests of the new system showed an improvement of 2.2 ± 1.5 % in capacity and 2.9 ± 1.5 % in COP compared to the performance of the original system, which was optimized by the manufacturer using conventional methods prior to the onset of this study.