A unified system to manage and scale AI workloads across diverse compute environments.
SkyPilot lets you run, manage, and scale your AI jobs from a single interface. It abstracts away the complexities of various infrastructures, including Kubernetes, Slurm, and over 20 cloud providers, allowing you to deploy models and orchestrate training with ease. You can start by defining your job requirements in a `sky.yaml` file and then use simple commands like `sky launch` to get going.
A unified system to manage and scale AI workloads across diverse compute environments.
AI developers and researchers who need to efficiently utilize and manage distributed AI compute resources.