Ray

Ray is the open-source AI Compute Engine: a unified, Python-native framework for scaling AI and machine learning workloads from a laptop to a cloud cluster.

Ray is your go-to for scaling Python applications, especially AI/ML. It abstracts distributed systems complexity, letting developers use a simple API (tasks, actors) to parallelize code across a cluster. The core runtime, Ray Core, powers a suite of specialized AI Libraries: Ray Data for scalable ingestion, Ray Train for distributed training, Ray Tune for hyperparameter optimization, Ray Serve for model deployment, and RLlib for reinforcement learning. This unified architecture enables seamless scaling of compute-intensive tasks, including large language models (LLMs) and GenAI, across heterogeneous compute (CPUs, GPUs) on any major cloud (AWS, GCP, Azure).

https://www.ray.io/