About this webinar
In MLOps, experimentation is the process of testing and comparing different ML models in production. It is an essential part of the continuous improvement of deployed models to make sure you are consistently getting the best out of your ML solutions in practice. Seldon’s expert team is here to help you get there faster. In this session, Paul Bridi, Product Manager will run through a practical session to demonstrate how Seldon can simplify the experimentation process. He’ll cover the principles of traffic splitting, A/B testing and shadow deployments, as well as showcasing Seldon’s experimentation offering in a demo.
Paul will show how easy it is to deploy and scale new model versions, as well as support real-time monitoring so teams can understand how their experiments are performing. He’ll also run an interactive Q&A to cover any questions the participants may have.
Speakers
Paul Bridi
What you'll learn
Understand experimentation with Seldon Core, including a demo on deploying and testing machine learning models using Kubernetes
Dive into key experimentation strategies covered were canary deployments, A/B testing, and shadow mode testing, illustrated with real-world use-cases
Look into technical challenges in experimentation such as routing, scaling, limiting downtime, and versioning were addressed, along with Seldon’s approach to solving these issues
Watch a demo of how Seldon Core’s functionality for running experiments, managing scaling and performance, and the upcoming model performance module