<p>This research presents novel approaches for optimizing the design of axial flow compressor blades. It introduces the utilization of machine-learning-aided techniques and surrogate models to enhance the performance of the optimization process. This study focuses on the rotary blade optimization in the transonic regime, specifically designed for a heavy-duty gas turbine. By employing a well-tuned parameterization technique that makes connections between control points in the spanwise direction having focused on the effective parameters in near-wall locations would bring improvements in streamlines, especially when passing the blade in near-wall positions. In this study, aerodynamic aspects of the flow are improved in parallel by having concentrated on the mechanical aspects of the blades. This is an important aspect when it comes to having safe structural performance parameters for both static and dynamic properties of the blade structure. Computational analyses using CFD-FEM simulations are conducted, considering mechanical constraints such as stress, force response, and flutter criteria before analyzing compressor mass flow and efficiency, respectively. The aero-mechanical objective functions and the blade's geometrical variables are related through a novel developed network. These objective functions encompass parameters such as mass flow rate, target stage efficiency, and total compressor efficiency over a semi compressor, which is a part of the main compressor. Having simulations, the optimum size of semi compressor is presented. The optimization algorithm was applied over the second stage of the compressor; findings demonstrate a significant improvement in the stage efficiency at the predefined stage pressure ratio, with having acceptable mechanical properties of the blades.</p>