Contribute Media
A thank you to everyone who makes this possible: Read More

Build Data Apps by Deploying ML Models as API Services


PyData SF 2016 Ramesh Sampath | Build Data Apps by Deploying ML Models as API Services

As data scientists, we love building models using IPython Notebooks / Scikit-Learn / Pandas eco-system. But integrating these models with an web app can be a challenge. In this tutorial, we will take our machine learning models and make them available as APIs for use by Web and Mobile Apps. We will also build a simple webapp that uses our prediction service.

Deploy your ML Models as a Service

In this talk, we will learn one way to take our Machine Learning models and make them available as a Prediction Service. We will work through the following steps.

Create a Simple Machine learning Model using Scikit-Learn / Pandas Pickle the model Using Tornado Web App, Make this model available as an API Service Build an Web App that uses this deployed Model Add Authentication to our Prediction API Optionally, add Redis to Cache Prediction Results Deploy the model in the Cloud (AWS) Please have Anaconda or Miniconda installed on your local machine. I will mostly be using Python 3.5, but Python 2.7 should be fine as well.


Improve this page