Soft Magnetic Composites (SMCs) possess promising electromagnetic characteristics and attract intense research and application interest in the engineering community. Fabrication of composites with customized architecture is feasible due to the recent advances in additive manufacturing techniques. The systematic progress of computational optimization has opened up the possibility of devising such structures. This article aims to optimally design the shape of inclusion in SMCs. Their advantages are potentially high magnetic permeability and low eddy current losses at a certain volume fraction of the iron material. The shape derivative is calculated for the descent direction and the level-set method is used to evolve the domain. We implement the algorithm in two-dimensional space considering the linear magnetic behavior of iron. In this article, the objective is the effective magnetic permeability at a given volume fraction of inclusion for magnetostatics.