Seldon.io

A talk at Spark+AI Summit: https://www.youtube.com/watch?v=D6eSfd9w9eA

Working with Bigmama last couple of years I did noticed that one of the main problems of machine learning projects is Deployement (As it is for Software Engineering Projects). Deployement means: make the models accessible to end users.

Many solutions have approached that specific problem: mlflow, corext-ai, seldon.io

We'll try to explain what makes seldon.io a good solution for machine learning deployments.