Union.ai Closes $38.1M Series A for MLOps
AI development infrastructure company Union.ai has closed a $38.1 million Series A funding round. The company, which builds on the open-source Flyte workflow orchestration platform, aims to simplify and scale ML pipelines in Kubernetes environments. The investment is targeted at improving reproducibility and developer velocity for AI teams.
The Series A was led by existing investor New Enterprise Associates (NEA), with participation from Nava Ventures and new investor Mozilla Ventures. This round includes a previously announced $19.1 million portion raised in 2023. Union.ai was co-founded by CEO Ketan Umare, who previously helped create Flyte while working as an engineer at Lyft. Before his time at Lyft, Umare worked at both Amazon and Oracle, building cloud infrastructure products. The Flyte project originated inside Lyft in late 2016 to handle complex ML challenges like calculating driver ETAs, which involved training multiple models on large datasets. Lyft open-sourced Flyte in 2020, and by mid-year, it was already powering over a million pipelines internally. Flyte's architecture is Kubernetes-native, designed to orchestrate tasks in isolated containers, which allows for dynamic resource allocation at runtime. This means engineers can define resource needs like GPUs directly in their Python code without changing the underlying infrastructure. The new capital is intended to support the commercial launch of Union 2.0 and the continued open-source development of Flyte 2. Planned features for Flyte 2 include pure Python authoring, improved debugging, and crash-resilient workflows. The open-source version of Flyte has surpassed 80 million downloads, while Union.ai's data validation framework, Pandera, has over 100 million downloads. Current enterprise customers using the platform include Spotify, Carfax, Hopper, and HederaDx.
