Build an Event-Driven Machine Learning Pipeline on Kubernetes

Explore the construction of an event-driven machine learning pipeline on Kubernetes in this 35-minute conference talk by Animesh Singh and Hou Gang from IBM. Delve into the growing importance of AIOps and the need for transparent AI pipelines that can detect data drift, bias, and model accuracy loss. Learn how to leverage Kubernetes components such as build, eventing, serving, and pipelines to create an end-to-end AI pipeline capable of detecting events, notifying users, and taking actions. Discover techniques for building and running data scientists' code, performing A/B testing, implementing dark launches, and orchestrating the entire workflow from model training to operations. Focus on eventing and pipeline CRDs from Kubernetes to understand how this process can be automated, enhancing the efficiency and reliability of machine learning pipelines in a Kubernetes environment.