The modern AI models are larger and more complex than ever. Enterprises are looking at implementing flexible and agile architectures to accommodate those requirements. It means hybrid and multi-cloud architectures. The challenge with distributed environments is the handling of data. There is data gravity and the difficult task of orchestrating the data across the various environments. How to address security and cost?
In this episode, our guest is Bin Fan, a founding member of Alluxio. The company provides a data orchestration platform that brings your data closer to computing across clusters, regions, clouds, and countries.
Bin outlines the challenges of dealing with data across distributed systems and the amount of data. Additionally, Bin discusses what kind of performance issues can be observed in ML workloads.
Join us in the podcast and hear more about the challenges of ML training data.
A link to this podcast “Utilizing AI” episode can be found here, or click on the video below.