Union.ai
by Union.ai
The commercial platform for Flyte — build durable, dynamic AI workflows and agents in pure Python
Union.ai is an AI development and orchestration platform, built by the team behind the open-source Flyte orchestrator, that lets engineering teams author durable, dynamic AI and ML workflows in pure Python and run them at scale. It is built for data and ML engineering teams productionizing training, inference, and agentic AI pipelines.
Union.ai, headquartered in Bellevue, Washington and founded in 2020 as the enterprise platform for the open-source Flyte orchestrator (over 80 million downloads), is led by CEO and co-founder Ketan Umare, part of the team that originally created Flyte at Lyft. The platform unifies data, models, and compute — spanning orchestration, model training, inference, and observability — to help teams move from experimentation to production while handling the non-deterministic, long-running nature of AI workloads. Its Union 2.0 release (Flyte 2) adds pure-Python authoring so engineers can write, test, and version workflows locally then run them at scale in the cloud, live remote debugging with failure logs, dynamic runtime decision-making for agentic AI, automatic infrastructure provisioning for long-running workflows, and crash-resilient pipelines with built-in caching and retries that resume after failure instead of restarting. Union.ai says more than 3,500 companies use infrastructure powered by its stack, and it also maintains Pandera, a data-testing framework with over 100 million downloads. In February 2026 the company closed a $38.1M Series A led by existing investor NEA, with Nava Ventures and new investor Mozilla Ventures.
At a Glance
- Category
- Infrastructure & Cloud
- Pricing
- Contact for pricing, Free
- Target Market
- Data Scientists, ML Engineers, Enterprise Developers, CTOs
- Founded
- 2020
- Headquarters
- Bellevue, United States
- Customers
- More than 3,500 companies use infrastructure powered by Union.ai
Key Features
- ✓Pure-Python authoring
Write, test, and version workflows locally in pure Python, then run them at scale in the cloud.
- ✓Dynamic runtime decisions
Fully dynamic workflows that make runtime decisions, suited to agentic AI applications.
- ✓Crash-resilient pipelines
Workflows resume after failure using built-in caching and retries instead of restarting from scratch.
- ✓Live remote debugging
Debug running workflows remotely with detailed failure logs.
- ✓Built on open-source Flyte
Enterprise platform for Flyte, the open-source AI orchestrator with over 80 million downloads.
Capabilities
Use Cases
- •Production ML training and inference
Move models from experimentation to reliable, scalable production pipelines.
- •Agentic AI workflows
Build dynamic, durable agent workflows that adapt at runtime and recover from failures.
- •Data and pipeline reliability
Reduce time spent maintaining brittle pipelines with caching, retries, and observability.
Ideal For
Best For
- ✓Orchestrating durable AI and ML pipelines in pure Python
- ✓Productionizing model training and inference at scale
- ✓Building dynamic, crash-resilient agentic AI workflows
Integrations
Market & Ratings
More than 3,500 companies use infrastructure powered by Union.ai
Market Analysis
Pros
- ✓Built on widely adopted open-source Flyte
- ✓Pure-Python developer experience
- ✓Durable, crash-resilient workflows for agentic AI
Cons
- ✗Managed platform pricing not public
- ✗Primarily aimed at engineering teams rather than business users
Pricing
Open source (Flyte)
$0
- ✓Open-source Flyte orchestrator
- ✓Self-hosted
Union.ai platform
Contact for pricing
- ✓Managed orchestration
- ✓Union 2.0 features
- ✓Enterprise support
The underlying Flyte orchestrator is free and open source; the managed Union.ai platform is enterprise and priced via sales.
Sources
This page was written from 2 sources, 1 on domains other than union.ai.
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
MinIO AIStor
S3-compatible object store rebuilt as the memory, table and object foundation for enterprise AI
Daytona
Sub-90ms stateful sandboxes that give every AI agent its own disposable computer
Volta
AI factories financed, built and operated like a utility — long-term GPU capacity for labs without hyperscaler balance sheets
Redis Iris
Real-time context engine giving AI agents governed retrieval, live operational data and durable memory