Dask
Focuses on Python parallelism but with less emphasis on AI/ML-specific workflows compared to Ray.
Kubeflow
Provides machine learning workflow orchestration but is more Kubernetes-centric than Ray.
Horovod
Specializes in distributed training of deep learning models, whereas Ray supports broader distributed Python workloads.
Flyte
Focuses on workflow orchestration for ML pipelines, complementing Ray's distributed compute capabilities.