The Sandia National Laboratories mission requires risk-informed decision making around complex engineering and science phenomena. Our reliance on computational models for insight in this process demands they have a certificate of credibility, including verification, validation, and quantification of uncertainties. I will motivate the uncertainty quantification (UQ) process and then survey UQ algorithms and software developed at Sandia to address typical computational engineering challenges such as simulation cost, nonlinearity, non-smoothness, and potentially large parameter spaces. Emerging reliability, stochastic expansion, and interval estimation algorithms in our Dakota software address these challenges. These can be employed in mixed deterministic/probabilistic analyses such as optimization under uncertainty. Application examples will include nuclear reactor performance assessment and micro-electro-mechanical system (MEMS) design. Time permitting, I will highlight some problems in multi-physics and complexity that lead to UQ research challenges.