This article presents a soft computing methodology to design turbomachinery components experiencing strong shock interactions. The study targets a reduction of unsteady phenomena using evolutionary optimization with robust, high fidelity, and low computational cost evaluations. A differential evolution (DE) algorithm is applied to optimize the transonic vane of a high-pressure turbine. The vane design candidates are examined by a cost-effective Reynolds-averaged Navier–Stokes (RANS) solver, computing the downstream pressure distortion and aerodynamic efficiency. A reduction up to 55% of the strength of the shock waves propagating downstream of the stand-alone vane was obtained. Subsequently to the vane optimization, unsteady computations of the vane–rotor interaction were performed using a non-linear harmonic (NLH) method. Attenuation above 60% of the unsteady forcing on the rotor (downstream of the optimal vane) was observed, with no stage-efficiency abatement. These results show the effectiveness of the proposed soft optimization to improve unsteady performance in modern turbomachinery exposed to strong shock interactions.