 This study shows that machine learning models based on the one electron reduced density matrix can be used to generate surrogate electronic structure methods for systems ranging from small molecules to complex compounds. These surrogates can predict molecular observables, energies and atomic forces using standard quantum chemistry or a second machine learning model without the need for computationally expensive algorithms such as self-consistent field theory. The algorithms are packaged in an efficient and easy to use Python code, QM learn, accessible on popular platforms. This article was authored by Shu Etching Xiao, Lucas Payto, Mark E. Tuckerman and others.