Containerization is the favoured method of working with applications on the cloud. When Data Science walks into the party, with its history of non-container workloads and humans who don't know containers, things turn chaotic. In this session, we cover the different taxonomies of AI/ML workloads and how best to make containers out of them. The responsibility just does not stop there for the MLOps persona - if that's you. The entire pipeline has to be designed in a manner that supports could native technologies that adhere to the expectations of a modern distributed system.

Top takeaways will include:

• Designing an AI/ML pipeline within the cloud native ecosystem
• The right level of abstraction for data scientists
• How not to lose focus on software supply chain security