First-principles descriptions of point defects in semiconductors and oxides has been a long-standing challenge.  Recently, we have been exploring the application of quantum Monte Carlo methods to predictive modeling of defect energetics, thermal transition levels, and optical ionization energies.  Quantum Monte Carlo methods are a suite of stochastic tools for solution of the many-body Schrodinger equation.  Due to its direct treatment of electron correlation, the QMC method offers the possibility of parameter-free, high-accuracy, systematically improvable analysis.  In this talk, I will present some results from our QMC descriptions of doped oxides including manganese oxide and zinc oxide, and as well as some semiconductors of interest for thin-film photovoltaics.  The use of a many-body approach gives insights to the challenges underlying accurate descriptions of phase stabilities, doping, etc. in these systems.  Additionally, statistical analysis of the many-body wave functions from QMC provides some diagnostic assessments to reveal the physics that may be missing from other first-principles simulation approaches.