This document provides an agenda and overview for a meetup about using the machine learning platform H2O. H2O can handle large datasets with billions of rows and hundreds of gigabytes of data. It performs computations very quickly, near Fortran speeds. The meetup will demonstrate using H2O with the MNIST handwritten digit recognition dataset and a bodily injury claims dataset from Allstate. For the MNIST data, random forest classification models will be built and the class label counts inspected. For the Allstate data, generalized linear models will be demonstrated for supervised prediction tasks.