In this chalk talk, we discuss building models using Amazon SageMaker and deploying them on to AWS Snowball Edge devices. Then we walk through how you can run analytics on the edge using Jupyter Notebooks to explore the raw data and model output data. Snowball Edge now supports Amazon EC2 and has 90 TB of storage plus 750 GB SSD attached to a virtual machine equivalent to an m4.4xlarge instance. Running machine learning (ML) models on Snowball makes it easy to process and store data in the field, and allows you to explore the data in real time without cloud connectivity. Once you have connectivity restored, you can upload the data to Amazon S3 for advance analytics and data exploration.