This article illustrates the opportunities that combining computational modeling and systematic design optimization techniques offer to facilitate the design process of shape memory alloy (SMA) structures. Focus is on shape memory behavior due to the R-phase transformation in Ni-Ti, for which a dedicated constitutive model is formulated. In this paper, efficient topology and shape optimization procedures for the design of SMA devices are described. In order to achieve fast convergence to optimized designs, sensitivity information is computed to allow the use of gradient-based optimization algorithms. The effectiveness of the various optimization procedures is illustrated by numerical examples, including the design of a miniature SMA gripper and a steerable SMA active catheter. It is shown that design optimization enables designers of SMA structures to systematically enhance the performance of SMA devices for a variety of applications.