Get Started With NVIDIA Run:ai

NVIDIA Run:ai is a GPU orchestration and optimization platform that accelerates AI operations by dynamically scheduling, allocating, and managing GPU resources for AI workloads. It helps organizations maximize GPU utilization, scale both training and inference workloads efficiently, and integrate seamlessly into hybrid or multi-cloud AI infrastructure with minimal manual effort.

NVIDIA Run:ai supports the entire AI lifecycle—from data processing and distributed training to inference workloads—enabling dynamic orchestration and scaling of complex machine learning jobs across distributed GPU clusters. It can handle interactive sessions, batch training jobs, and ongoing inference jobs, helping teams run more workloads in parallel with higher utilization.

NVIDIA Run:ai is built on top of Kubernetes, extending its capabilities with an advanced AI scheduler that automates GPU resource allocation, workload submission, sharing, and scheduling. This enables users to run AI workloads within a familiar Kubernetes ecosystem while benefiting from intelligent GPU orchestration.

NVIDIA Run:ai helps organizations:

• Maximize GPU utilization by pooling GPU resources and dynamically assigning them to workloads.
• Reduce idle compute capacity and infrastructure costs by running more jobs in parallel.
• Support hybrid deployments across on-premises, cloud, or multi-cloud environments with centralized management and visibility.
• Integrate with existing AI tools and frameworks, thanks to its open, API-first architecture.

When you log into NVIDIA Run:ai for the first time, guided onboarding flows help you get started quickly:

• Administrators are walked through installing the cluster, configuring single sign on, and inviting teams.
• Researchers are guided through creating their initial workspace and starting their first workloads directly from the UI.