Deploying Machine Learning Models as REST APIs

Presenter: Paige Bailey

10:40 am - 11:05 am

ai, machine learning, devops, api use cases, rest, api best practices

You’ve sourced your data, then cleaned and structured it. You’ve used that data to train and test many algorithms and model topologies; and you finally have one that meets your business requirements for accuracy. Now what?

In this session, we will explore best practices for deploying a machine learning model (Keras + TensorFlow) as a REST API, using Python Flask and a cloud-based infrastructure. We will define a schema for inputs and outputs; a script to initialize and run the model; and a serverless function to trigger it.

By the end of this session, you will know how to build a production-ready endpoint for your machine learning model that any software engineer in your organization can connect to: no more static CSV files or business reports.