Tecton.ai, a company launched by three former Uber engineers with the aim of making the machine learning feature store concept widely accessible, has revealed a $35 million Series B funding round. This announcement comes only seven months after the company publicized its $20 million Series A financing.

During a conversation with the company in April, they were collaborating with initial clients during a beta phase of their product. Today, alongside the funding news, they are also announcing the full public release of their platform.

Similar to their Series A round, this investment is being jointly led by Andreessen Horowitz and Sequoia Capital. The company’s total funding now reaches $60 million.

The significant commitment from these two investment firms stems from the particular challenge in machine learning that Tecton is addressing. According to company CEO and co-founder Mike Del Balso, “Our focus is on enabling organizations to deploy machine learning into real-world applications. Our entire company is dedicated to assisting someone in creating a functioning machine learning application – one that powers their fraud detection system or another critical function – and simplifying the process of building, deploying, and maintaining it.”

They accomplish this through the implementation of a feature store, a concept they pioneered and which is rapidly evolving into a distinct category within machine learning. The recent announcement of the Sagemaker Feature Store by AWS was viewed by the company as substantial confirmation of their approach.

Tecton describes a feature store as a comprehensive machine learning management system. This system encompasses the data transformation pipelines that generate feature values, the storage and management of all feature data, and the delivery of a consistent data set.

Del Balso explains that this system integrates seamlessly with other components of a machine learning stack. “When developing a machine learning application, you typically utilize a machine learning stack that may include a model training system, a model serving system, or an MLOps layer for model management. Alongside these, you have a feature management layer – the feature store, which is where we come in – providing an end-to-end lifecycle for your data pipelines,” he stated.

Backed by substantial investment, the company is experiencing rapid growth, increasing its headcount from 17 to 26 employees since April, with plans to more than double that number by the end of the following year. Del Balso and his co-founders are dedicated to fostering a diverse and inclusive workplace, acknowledging the challenges involved in achieving this goal.

“We’ve made this a key focus of our recruitment efforts. It’s a difficult undertaking that requires significant dedication; it cannot be treated as a secondary priority,” he said. To this end, the company has actively participated in and sponsored diversity hiring events, concentrating its recruitment efforts on identifying a broad range of candidates.

In contrast to many startups, Del Balso expresses a desire to return to a traditional office environment as soon as possible, believing it will strengthen interpersonal relationships among employees.