Discovery of new materials with interesting properties is traditionally driven by intuition, chemical substitutions, and serendipitous discoveries. Recently, technology giants such as Google and Microsoft are using large databases, generative machine learning, and autonomous labs to accelerate this process, but the experimental realization of the computational predictions often fail because disorder is widespread in real materials. In this project, the student will predict synthesizable and disordered materials using model parameters from density functional theory (DFT). The target materials are multivalent high entropy oxides (HEOs). HEOs are heavily disordered yet crystalline oxides where five or more cations share one lattice site in roughly equal amounts. Oxides with heterogeneous cation valences can be realized as long as the total charge balance is preserved.