Deploying a model with Amazon Elastic Inference