ClearML
Open-source MLOps platform for experiments, pipelines, and model management
Durable AI and ML orchestration built on Flyte 2
Union.ai is an MLOps orchestration platform built on Flyte 2, a durable, infrastructure-aware open-source runtime for AI and ML workflows authored in pure Python, with enterprise features on top.
Union.ai develops Flyte, a widely used open-source orchestration engine for machine learning and data workflows, and offers a managed enterprise platform on top of it. In August 2026 the company announced the general availability of Flyte 2, a ground-up rebuild designed for modern AI, ML, and agentic workloads. Flyte 2 introduces what Union.ai calls the first durable AI runtime in open source: it provides fault-tolerant, resource-aware execution that understands and adapts the infrastructure and compute beneath the code, so workflows recover automatically from failures and scale to production. Developers author workflows in real Python without learning a new DSL, testing and versioning locally before running at scale in the cloud. Union.ai's commercial platform is built on the same core engine and adds enterprise capabilities such as real-time inference, observability, single sign-on, and role-based access control. This combination targets ML and AI engineering teams that need production-grade, durable orchestration with the flexibility of open source underneath.
Union.ai is an MLOps orchestration platform built on the open-source Flyte engine, whose Flyte 2 release delivers a durable, infrastructure-aware runtime for AI and ML workflows in pure Python.
Union.ai is the company behind Flyte, an open-source workflow orchestration engine widely used for machine learning and data pipelines. It offers a managed enterprise platform on the same core technology.
In 2026 Union.ai shipped Flyte 2 to general availability, rebuilding the engine for modern AI, ML, and agentic workloads and introducing a durable AI runtime in open source. This positions the company at the intersection of MLOps and agentic infrastructure.
Flyte 2 provides fault-tolerant, resource-aware execution that adapts infrastructure beneath the code, with workflows authored in pure Python and tested locally. It recovers automatically from failures and scales to production.
Union.ai's platform builds on this engine with real-time inference, observability, SSO, and RBAC for enterprise teams. Users can self-host Flyte or adopt the managed platform depending on their needs.
Union.ai targets ML engineers, data platform teams, and production ML organizations that need durable, governed orchestration. It is less suited to non-technical users, trivial jobs, or non-Python ML stacks.
ML and data engineers authoring and running pipelines.
Heads of ML platform and data engineering choosing orchestration.
MLOps and infrastructure engineers evaluating durability and scale.
Python-based ML organizations running production pipelines that need fault tolerance, scale, and enterprise governance.
Verify current funding details with the vendor or public sources.
Union.ai develops the open-source Flyte engine and offers a managed enterprise platform built on the same core.
Flyte 2 is a ground-up rebuild providing a durable, infrastructure-aware runtime that recovers from failures and scales AI/ML workloads, reaching GA in 2026.
No. You author workflows in pure Python, testing and versioning locally before running at scale.
Yes. Flyte is open source and self-hostable, while Union.ai adds managed enterprise capabilities.
It adds real-time inference, observability, single sign-on, and role-based access control on the same core engine.
Side-by-side pages for pricing, features, and best-fit use cases.
Open-source MLOps platform for experiments, pipelines, and model management
Open-source library for fast, memory-efficient LLM fine-tuning
Open-source framework that makes LLM fine-tuning reproducible from a single YAML config
Platform for fine-tuning and serving open-source LLMs (now part of Rubrik)