Abstract: When it comes to MLOps, storage and data are related — but far from the same. So why is a storage company doing yet another MLOps talk at GTC this year? We're here to help you focus on data and not think about storage. We're going to do this in two ways: First, we'll show you how storage can get out of the way of data science. Second, we'll show you how a modern data experience works to streamline machine learning and inference operations. It's important to think about how to scale simplicity and performance across all the various components of an AI infrastructures. The right solution gets the storage out of the way of the data science, and allows data scientists to focus on the DATA. It also simplifies life for the Infrastructure team, enabling simple, trouble-free operation and automation that moves toward Infrastructure-as-Code.
The entire endeavor becomes simpler when unifying around a single storage platform to safely hold all the data you need, scale as those needs change, and reduce time wasted copying data across pipeline steps and silos. The combination results in faster time to insights and more models quickly getting into Production. Check out this short talk to learn more.
Bio: Miroslav is a Field Solution Evangelist focused on AI and Analytics. He has extensive experience in storage systems, workload modeling, system architecture, tuning, and benchmarking. Miroslav loves both learning and teaching, and has translated those passions into helping organizations get value from innovative technologies. He has previously worked with top companies (including Facebook, NetApp, and Sun) and a number of startups, and is proud to share his experience. Much of Miroslav’s career has revolved around technical evangelism and performance — focused on system architecture and storage, and blending time with customers with time in the lab. He’s been at this for over 30 years now, with his first paid computer gig involving DOS batch scripts the summer before college.