Infrastructure
we will use AWS ML to start (EMR/Spark, S3, AWS ES, Sagemaker, AWS ML, AWS DeepLens, AWS Deep Learning AMI (TensorFlow, Gluon, Apache MXNet, Caffe2, Keras, PyTorch)

In particular sagemaker
https://console.aws.amazon.com/sagemaker/home?region=us-east-1#/landing

https://docs.aws.amazon.com/solutions/latest/data-lake-solution/deployment.html
http://docs.awssolutionsbuilder.com/data-lake/