In this section we will train a simple digit recognition model using the MNIST dataset provided through the TensorFlow library². To load the dataset, in the first cell of you notebook type the following. This will display the shape of the training data inputs (tx) and target (ty), and validation data inputs (vx) and target (vy). Let's now display a set of 10 sample images for each digit to understand how the data looks like. This will generate the following grid of images.

TensorFlow is an open-source machine learning (ML) library widely used to develop neural networks and ML models. Those models are usually trained on multiple GPU instances to speed up training, resulting in expensive training time and model sizes up to a few gigabytes. After they're trained, these models are deployed in production to produce inferences. They can be synchronous, asynchronous, or batch-based workloads. Those endpoints need to be highly scalable and resilient in order to process from zero to millions of requests.