GTC 2020: Accelerate ML Lifecycle with Containers, Kubernetes and NVIDIA GPUs (Presented by Red Hat)

Tushar Katarki, Red Hat | Zvonko Kaiser, Red Hat
GTC 2020
Data scientists desire a self-service, cloud-like experience to easily access ML modeling tools, data, and GPUs to rapidly build, scale, reproduce, and share ML modeling results with peers and developers. Containers and kubernetes platforms, integrated with NVIDIA GPUs, provide these capabilities to accelerate the training, testing, and deploying the ML models in production quickly. We'll provide an overview of how these technologies are helping solve real-world customer challenges. We'll walk through the various customer use cases and solutions associated with the combination of these technologies. We'll review the key capabilities required in a containers and kubernetes platform to help data scientists easily use technologies (such as Jupyter Notebooks, ML frameworks, etc.) to innovate faster. Finally, we'll share the platform options (for example, Red Hat OpenShift, Kubeflow), and examples of how data scientists are accelerating ML initiatives with containers and Kubernetes.