An R package efficiently implementing the algorithms in the Almost-Matching-Exactly framework for interpretable matching in causal inference. These algorithms match units via a learned, weighted Hamming distance that determines which covariates are more important to match on.