Arrikto, a company focused on accelerating the machine learning development process by enabling data scientists and engineers to handle data as code, is officially launching today and has secured $10 million in Series A funding. Unusual Ventures spearheaded the investment, and Unusual’s John Vrionis will be joining the company’s board.

“The technology developed at Arrikto assists organizations in navigating the challenges associated with implementing and maintaining machine learning applications,” explained Constantinos Venetsanopoulos, CEO and co-founder of Arrikto. “We significantly streamline the creation of complete machine learning pipelines. Specifically, we simplify the process of constructing, training, and deploying ML models to production environments utilizing Kubernetes, while also intelligently overseeing the associated data management.”

Similar to many platforms geared towards developers today, Arrikto emphasizes a “shift left” approach. The company contends that currently, machine learning teams and development teams often operate with differing methodologies and utilize distinct tools for model building and production deployment.

“In a manner analogous to how DevOps brought deployment earlier into the software development lifecycle for developers, Arrikto brings deployment earlier into the machine learning lifecycle for data scientists,” Venetsanopoulos clarified.

Arrikto also strives to lower the technical hurdles that continue to make machine learning implementation difficult for many businesses. Venetsanopoulos pointed out that, just as Kubernetes demonstrated the potential for a straightforward and scalable infrastructure, Arrikto can illustrate what a more manageable ML production pipeline could look like – and achieve this within a Kubernetes-native framework.

Kubeflow, the open-source machine learning toolkit for Kubernetes incubated by Google, forms the foundation of Arrikto’s offerings, and the company can be viewed as providing an enterprise-level version of Kubeflow. The team has also developed projects like MiniKF, which allows Kubeflow to run on a laptop, and Kale, which enables engineers to construct Kubeflow pipelines directly from their JupyterLab notebooks.

According to Venetsanopoulos, Arrikto’s technology delivers three key benefits: it simplifies the deployment and management of Kubeflow, empowers data scientists to utilize their existing tools, and establishes a portable data science environment that supports data versioning and sharing across teams and cloud platforms.

Since its initial launch from Athens, Greece in 2015, Arrikto has operated discreetly, yet the founding team of Venetsanopoulos and CTO Vangelis Koukis have successfully onboarded a number of major enterprises onto its platform. Currently, Arrikto serves over 100 customers, and while the company is not yet at liberty to disclose their identities, Venetsanopoulos mentioned that they include a prominent oil and gas corporation.

Despite Athens not being traditionally recognized as a startup center, Venetsanopoulos believes this is evolving, with a wealth of untapped talent available (and the company is also leveraging the funding to expand its sales and marketing presence in Silicon Valley). “There’s a high caliber of talent emerging from leading universities that remains largely unexplored, giving us a considerable advantage,” he stated.