TL;DR:
- Union AI secures $19.1 million in Series A funding from NEA and Nava Ventures.
- The company introduces the fully managed Union Cloud service for AI and data workflows.
- Flyte, an open source tool, enables seamless workflow automation for data, machine learning, and analytics.
- Union AI addresses infrastructure challenges faced by machine learning teams.
- Collaboration with Lyft led to the development of Flyte and the founding of Union AI.
- Flyte is adopted by notable companies like blackshark.ai, HBO, Intel, LinkedIn, Spotify, Stripe, Wolt, and ZipRecruiter.
- Real-world success stories demonstrate increased productivity and frequent model releases.
- Union AI extends beyond Flyte with Pandera for data testing and Union ML for model deployment.
- Union Cloud combines all elements, providing enterprise tools and ensuring data control and security.
- Greg Papadopoulos emphasizes the importance of maintaining control and ownership of data when leveraging large language models.
Main AI News:
Union AI, a promising startup hailing from Bellevue, Washington, has successfully secured a substantial $19.1 million Series A funding round from renowned investors NEA and Nava Ventures. This impressive investment comes in tandem with the company’s exciting announcement of the general availability of its fully managed Union Cloud service. These developments highlight Union AI’s unwavering commitment to revolutionizing the way businesses construct and coordinate their AI and data workflows through the power of a cutting-edge cloud-native automation platform.
Central to Union AI’s groundbreaking technology is Flyte, an exceptional open source tool meticulously crafted to facilitate the creation of top-tier workflow automation platforms specifically tailored to excel in the realms of data, machine learning, and analytics. By consolidating an assortment of critical functionalities into a single, unified platform, teams gain the ability to effortlessly construct their ETL pipelines, analytics workflows, and machine learning pipelines. While other projects may offer similar orchestration capabilities, Union AI differentiates itself by dedicating its efforts solely to the unique requirements of machine learning teams, ensuring optimal performance and efficiency.
Flyte originated within the esteemed corridors of Lyft, where the visionary CEO and co-founder of Union AI, Ketan Umare, spearheaded the development of the company’s earliest machine learning-based ETA and traffic models back in 2016. During this period, Lyft grappled with the arduous task of cobbling together disparate open source systems to effectively deploy these models into production environments.
Umare candidly acknowledged, “We got something running, but behind the scenes, it was a man behind the curtain. It was happening, but it was a lot of work.” It quickly became apparent that the challenges experienced by his team were by no means isolated incidents, as numerous departments within Lyft, including sizable teams, encountered similar struggles.
This inability to deliver effectively had severe consequences, with talented individuals departing the organization due to these obstacles. Umare recognized that the root cause lay in an infrastructure problem, impeding teams from realizing their full potential.
Determined to address this critical issue, Umare assembled a dedicated team and embarked on a mission to develop robust infrastructure tools capable of streamlining the model creation and production processes for these teams. However, tensions between software engineers and machine learning specialists arose due to the inherent disparities between traditional software development and the intricate nature of AI systems.
Umare passionately expressed, “The reason was that—I have distilled it—I think software and machine learning systems or AI products are inherently different beasts.” While software typically undergoes gradual refinement over time, AI models often experience degradation. Moreover, these models frequently undergo changes influenced by external factors beyond the users’ control. Consequently, utilizing the same infrastructure employed for traditional software deployments proves inadequate in the realm of AI.
In an inspiring turn of events, the dedicated team behind Flyte made the decision to embrace the open-source ethos and collaborate with like-minded individuals to cultivate a truly machine-learning-native platform. Recognizing the immense potential of their work, Ketan Umare and four other esteemed members of the original Flyte team embarked on a remarkable journey, establishing a startup that would solidify their vision. Thus, in late 2020, Union AI was born, propelled by the core ideas and principles underpinning the Flyte open-source project.
Today, Flyte stands as a testament to its groundbreaking capabilities, as it finds itself adopted by an impressive array of industry leaders, including renowned companies such as blackshark.ai, HBO, Intel, LinkedIn, Spotify, Stripe, Wolt, and ZipRecruiter. Embracing these collaborations with notable industry giants has allowed Union AI to validate the effectiveness of its platform on a grand scale.
Ketan Umare eloquently expresses the significance of these partnerships, stating, “The fun thing about working with these large companies – what we do in the open source – is that we are working on some of the biggest models on our platform. So we know it works, and we didn’t have to build anything specifically because we’ve been doing this for years. We just had to extend a couple of things.”
Real-world success stories abound, with Mick Jermsurawong, a machine learning infrastructure engineer at Stripe, sharing his firsthand experience: “Based on a single team, we see 10x more offline training jobs dispatched from Flyte, and that results in 5x more frequent model releases with sizable business gains. I think the realization here is that ML productivity is not a nice-to-have but actually a business requirement.” Such testimonies serve as powerful evidence of the profound impact Union AI’s platform has on organizations, unlocking unparalleled productivity and propelling them toward tangible business growth.
Union AI, however, has set its sights beyond merely offering Flyte as a service. The visionary team has also introduced two groundbreaking frameworks: Pandera, a comprehensive data testing framework, and Union ML, a powerful framework that seamlessly integrates with Flyte, empowering teams to effortlessly build and deploy their models using their existing toolset.
To bring these remarkable components together and provide an all-encompassing solution, Union Cloud was crafted. This transformative platform seamlessly merges the various elements while augmenting them with a suite of enterprise tools, including the invaluable single sign-on feature, ensuring unparalleled convenience and security.
It is crucial to recognize that the adoption of machine learning, particularly large language models, inevitably raises concerns surrounding privacy and information security. Companies are becoming increasingly cautious about utilizing services that compromise their control over sensitive data.
Greg Papadopoulos, a venture partner at NEA, keenly observes the significance of this challenge and expresses his excitement for the strides made by the Union AI team: “Combining the power of big models with rich company data has to be handled with care – that’s one of the reasons why we’re so excited about the progress made by the Union.AI team, first with Flyte and now with Union Cloud. This is exactly what people are demanding and a real differentiator: Let me exploit the power of large language models while maintaining control and ownership of my data.”
Conlcusion:
Union AI’s significant Series A funding and the launch of their fully managed Union Cloud service underscore their position as a key player in revolutionizing AI and data workflows. Their innovative open source tool, Flyte, coupled with their dedication to addressing the unique needs of machine learning teams, has garnered adoption by renowned companies across various industries.
This success, combined with the introduction of frameworks like Pandera and Union ML, showcases Union AI’s commitment to driving productivity and business growth in the market. By prioritizing data control and security while harnessing the power of large language models, Union AI is poised to meet the growing demand for effective and privacy-conscious AI solutions, making a substantial impact on the market.
