We develop simulation frameworks that connect the atomic structure and composition of materials to key macroscopic properties, such as thermal conductivity, resistance to thermo-mechanical stresses, and neutron irradiation. The group leverages first-principles simulations and AI methods to solve these theories with quantitative accuracy—enabling understanding, design, and discovery of atomistic structures that yield targeted macroscopic material properties. This theory-driven approach dramatically accelerates and reduces the cost of discovering more durable plasma-facing materials for nuclear fusion reactors.