Gaussian measure has for decades been recognized as the appropriate measure to use in infinite-dimensional analysis, and calculus on such measure spaces has become a valuable tool in the analysis of stochastic processes and their applications. For infinite-dimensional curved spaces, the analogue of Gaussian measure is heat kernel measure. We'll discuss heat kernel measures in a special class of infinite-dimensional spaces and provide motivation for the construction. In particular, these spaces admit a natural hypoelliptic structure, and we're able to show smoothness results for heat kernel measures under both elliptic and hypoelliptic conditions. Parts of this talk are based on joint work with F. Baudoin, D. Dobbs, B. Driver, N. Eldredge, and M. Gordina.