Q&A: Choosing the right compute infrastructure for machine learning

Seemingly overnight, machine learning (ML) exited the world of aspirational technology and entered the mainstream. Organizations of every size and across nearly every industry want in on the action—and ML is realistically within reach for all because of the cloud. The cloud brings together data, low-cost storage, security, and ML services, along with high-performance CPU (central processing unit) and GPU (graphics processing unit) based instances for model training and deployment. Now, organizations can store as much data and have as much high-performance compute as they need elastically, so realizing the value of ML can happen much faster.

However, with the emergence of a wide breadth and depth of cloud infrastructure options and services, making the right selection for your use case can be difficult. Many executives are asking, “What factors should I consider when choosing the right cloud compute infrastructure and services for my ML objectives?”

For the answers to that question and more, we turned to Dr. Bratin Saha, VP of AI/ML services. Read on to discover guidance and best practices for evaluating the infrastructure requirements of your ML workloads—and for ensuring you make the right choices to meet those needs.