We propose a novel type of quantum heat engine based on the ultrafast dynamical control of the magnetic properties of a nano-scale working body. The working principle relies on nonlinear phononics, an example for dynamical materials design. We describe the general recipe for identifying candidate materials, and also propose Cr$_{2}$O$_{3}$ as a promising working body for a quantum Otto cycle. Using a spin Hamiltonian as a model for Cr$_{2}$O$_{3}$, we investigate the performance in terms of efficiency, output power, and quantum friction. To assess the assumptions underlying our effective spin Hamiltonian we also consider a working substance composed of several unit cells. We show that even without an implementation of transitionless driving, the quantum friction is very low compared to the total produced work and the energy cost of counterdiabatic driving is negligible. This is an advantage of the working substance, as experimentally hard-to-implement shortcuts to adiabaticity are not needed. Moreover, we discuss some remarkable thermodynamic features due to the quantumness of the proposed system such as a non-monotonic dependence of the efficiency on the temperature of the hot-bath. Finally, we explore the dependence of the performance on the system parameters for a generic model of this type of quantum heat engine and identify properties of the energy spectrum required for a well-performing quantum heat engine.