SkyPilot

SkyPilot is an open-source framework that lets you run AI and machine learning workloads across multiple cloud providers and GPU clusters with a single, unified interface.

Instead of learning the specific APIs, CLI tools, and pricing structures of each cloud platform, you simply define your job once and SkyPilot handles the rest — finding the best available hardware, launching the cluster, running your workload, and shutting everything down when it's done.