This is a proof of concept H2o3 AutoML integration with Amazon SageMaker. Most of the code is adapted from https://github.com/awslabs/amazon-sagemaker-examples/tree/master/advanced_functionality/scikit_bring_your_own/container and customized to support H2o3 AutoML. It's highly recommended that you read the docs on that repository first.
The training and testing data used in this example are from the UCI machine learning repository and are at https://archive.ics.uci.edu/ml/machine-learning-databases/adult/ .This dataset contains a bunch of features such as a person's age, marital status etc, and we're writing a binary classifier to determine whether the person's income is <= 50K, or if it's > 50K.