Abstract: The capabilities of intelligent applications often seem like magic to users, but the machine learning and artificial intelligence techniques that enable these features are more accessible than you might think. Developing intelligent features doesn’t require esoteric math or high-performance hardware, but it does require you to start with data rather than with code and to adapt your existing engineering practice to build and manage predictive models in addition to conventional software artifacts.
This hands-on tutorial will introduce machine learning workflows and concepts in the context of a concrete problem and show you how to integrate them into the application development work you’re already doing, focusing on the habits and processes that will help you to get meaningful results from predictive models. We’ll work through a case study of a real application and you’ll leave having learned
● How to process, transform, and visualize data
● How to train and evaluate predictive models
● How the same build and test pipelines that support your software engineering work also
enable putting machine learning into practice
● Potential pitfalls of incorporating machine learning into your application -- as well as how
to avoid them
● How contemporary cloud infrastructure dramatically streamlines developing application
You’ll do all of this with while building models to solve a real problem and publishing them as microservices, all on a completely open-source stack including OKD, Python, Jupyter notebooks, Numpy, Pandas, Scikit-Learn, and Vega.
Bio: William Benton leads a team of data scientists and engineers at Red Hat, where he has focused on enabling machine learning workflows and data processing pipelines in cloud-native environments while solving some fun problems with data.