This document discusses using MapReduce for exact Bayesian network inference. It provides background on Bayesian networks and their applications. It describes the CPCS network used for medical diagnosis that contains 422 nodes. Exact inference on this full network was previously impossible due to its large size, but MapReduce allows breaking the computation into smaller parallel pieces to solve it. The results show Bayesian network inference that previously took hours can now be done in minutes using Hadoop MapReduce.