Union.ai was founded in 2021 by the engineers who created Flyte™.
Union’s mission: to orchestrate the myriad processes required to train and run the most complex and dynamic machine learning and data products at scale.
During our time at Lyft, the co-founders of Union encountered numerous challenges faced by ML and data processing teams. In response to these challenges, we created Flyte™, a platform for orchestrating machine learning and data processing jobs. Initially, our small team launched Flyte™ in late 2016 to deliver models for a core product — estimated time of arrival (ETA).
In early 2020, we decided to open-source Flyte™, and it became the 25th hosted project of the Linux Foundation AI & Data. Soon after, the world experienced significant changes due to the pandemic, which impacted all organizations, including Lyft. Faced with a crucial decision, we had to choose between joining a different company to replicate our work or continuing the work we had started. We chose the latter, recognizing the struggles faced by many emerging ML product teams and the need to build a truly open ecosystem to help them.
Although starting a business was not part of our initial plan, we believed that creating a company dedicated to making ML and data teams agile and efficient was the best way forward. We were fortunate to connect with like-minded investors who shared our long-term vision and commitment to open-source ecosystems. Union.ai has become the driving force behind Flyte™, which has been adopted by over 500 organizations and production teams worldwide
In 2022, we welcomed Pandera into the Union.ai family. Pandera is a powerful tool designed to perform complex data validation tasks while maintaining a user-friendly API. It enables data scientists and engineers to define, validate, and document data schemas seamlessly, ensuring data consistency and quality throughout the pipeline.
By incorporating Pandera into the Union.ai ecosystem, we strive to offer our users a comprehensive set of tools to manage their data workflows more effectively. With Pandera's capabilities, we can tackle data quality challenges and ensure that ML models and data-driven insights are based on reliable and consistent information.
As the Flyte™ community grows, we have observed that many users encounter infrastructure challenges when setting up a Kubernetes-native platform. However, once they overcome these initial hurdles, users appreciate Flyte™’s scalability, efficiency and features. To facilitate the development of ML products, we aim to help teams get started with Flyte™ quickly and easily. We developed Union Cloud, a managed version of Flyte™, to simplify the adoption of Flyte™ for enterprises, mid-sized organizations, and even small teams.
Moving forward, we will continue to introduce features that benefit the entire community, such as global sharing, enhanced visualization, and services to improve Kubernetes deployments. Our dedication to simplifying data workflows and empowering efficient ML and data teams remains central to our mission.
We are global
We are a remote-first company, and you can find our employees in these locations:
Washington, Georgia, Wisconsin, Texas, Connecticut, Illinois, New York, California
China, India, Moldova, Australia, Germany, Switzerland, Colombia, Canada
Kelsey Hightower is a renowned technology advocate and thought leader in the cloud-native computing space. With extensive experience in Kubernetes, Go, and cloud technologies, Kelsey has made significant contributions to the open-source community. As a frequent conference speaker and author, he shares his expertise, inspiring developers and organizations to embrace innovative solutions and build scalable, reliable systems.
Wes McKinney is a software entrepreneur focused on data tools. He created Python’s pandas project and co-created Apache Arrow. He wrote “Python for Data Analysis” and serves on the PMC for Apache Parquet. He is now the CTO of Voltron Data, which develops accelerated computing technologies powered by Apache Arrow.
Matt Klein is a software engineer at Lyft and the creator of Envoy. For over 20 years across a variety of companies, he has worked on increasing the power and simplifying the operation of operating systems, virtualization, distributed systems and networking. Some highlights include leading the development of Twitter’s L7 edge proxy and working on high-performance computing and networking in Amazon’s EC2.