Maria Emelianenko George Mason University This talk will survey recent developments in two areas critical for advancing materials design, where defects play an essential role. First is mesoscale kinetic modeling of polycrystalline materials, focused on the task of understanding how statistical distributions develop in the process of coarsening of materials microstructure and how these distributions in turn relate to materials properties. The challenges here include the design of reliable benchmarks for curvature-driven growth, vertex and Monte Carlo grain growth simulation codes, as well as the development of coarse-grained kinetic theories capable of capturing realistic materials behavior. These and other questions will be discussed in the context of nonlocal evolution theory and particle gas dynamics, and sometimes unexpected connections with other fields of science will be revealed. The other part of the talk will be concerned with phase diagram calculation methods, where robust and accurate numerical optimization methods are required to prevent costly mistakes. A universal Gibbs energy minimization formulation will be discussed that allows to link traditional Calphad codes and databases to state-of-the-art optimization engines, paving the road to a more intelligent automated phase data exploration.