The Bayesian Observationally-constrained Statistical-physical Schemes (BOSS) are a family of bulk microphysics scheme developed using data-driven methods. Microphysical process rates are assumed to be related to the moments of the drop size distribution according to a simple ansatz (sums of power laws) with a number of free parameters. We choose values of these parameters using numerical Bayesian inference methods, specifically Markov-Chain Monte Carlo, which provides an estimate of parameteric uncertainty as well as “best-fit” parameter choices. We find that BOSS can very effectively emulate the behavior of the Tel Aviv University (TAU) bin microphysics scheme in a simple 1-D driver; accurate emulation requires using “online” rather than “offline” evaluation of the BOSS schemes’ behavior. Using three cloud moments is also needed to simulate the process-level physics with the highest level of accuracy. Our next steps include emulation of TAU in LES (i.e. with more realistic/complex clouds), incorporation of real-world observations, and evaluation of “single-category” versions of BOSS.