Awesome AI Agents › AI Workflow Orchestrators

flyteorg/flyte

⭐ 7483 Go added to this list on 2026-05-04 repository created 2019-10-21

Flyte is a robust, open-source orchestration engine tailored for machine learning workflows. It enables data scientists and engineers to define and manage complex AI pipelines, models, and agents in a scalable and resilient manner. The platform emphasizes the use of pure Python for defining workflows, allowing for intuitive and programmatic control over ML operations. Key features include dynamic task creation, robust error handling, data provenance tracking, and seamless integration with containerization technologies like Kubernetes. Flyte is particularly well-suited for orchestrating tasks related to data processing, model training, model serving, and integrating AI agents within larger systems. It supports local development workflows with a rich TUI and offers capabilities for deploying and serving models directly from the platform. While the full open-source backend for Flyte 2 is under development, it aims to provide a Kubernetes-native infrastructure for distributed, multi-node deployments. Flyte is a graduated project of the LF AI & Data Foundation, highlighting its maturity and community support.

https://github.com/flyteorg/flyte

mlopsworkflow orchestrationmachine learningAI agentspythonkubernetesdata scienceproductiondeclarativescalefine-tuningdata analysisllmgrpcgolangAI development toolsorchestration engine