Gaussian processes, characterised by their fully specified means and covariances, offer a flexible non-parametric approach to surrogate modelling, statistical learning, and uncertainty quantification.