image
image

2 min read


Home » Blog » Provectus Announces a New Partnership with Tecton to Collaborate on Feature Store for Machine Learning

Provectus Announces a New Partnership with Tecton to Collaborate on Feature Store for Machine Learning

Author:
Provectus, AI-first consultancy and solutions provider.

Provectus to contribute to bringing the next generation of Feast to AWS

PALO ALTO – May 18, 2021Provectus, an AI-first consultancy and solutions provider, today announced its partnership with Tecton, the enterprise feature store company, with the aim of standardizing and enhancing the machine learning production stack for enterprises.

Provectus and Tecton will closely collaborate in three major areas:

  1. Joint contributions and enhancement of Feast. Provectus will contribute to the next generation of Feast, the leading open-source feature store. While Tecton will remain the major contributor to Feast, Provectus’s efforts will be focused on bringing Feast to AWS as well as on core data models and APIs.
  2. Provectus — a partner of choice. Provectus will become a partner of choice for Tecton, providing users of both the Tecton and Feast feature stores with consulting and professional services.
  3. Design and standardization of open-source feature store APIs. Provectus and Tecton, along with users of ML feature stores, will join forces to define common, non-opinionated vendor agnostic APIs for feature stores.

Feature Stores for Machine Learning is a relatively new concept. They constitute the most critical piece in the modern ML production stack. As such, feature stores need to be clearly defined, standardized, and aligned with industry best practices. Provectus and Tectonhave the experience and expertise to accomplish those goals.

“When we introduce a feature store to our clients, the platform acts as an essential component that provides a clear path to production for online ML use cases such as Recommendations and Fraud Detection. You will have to build it one way or another,” says Stepan Pushkarev, CTO at Provectus.

Tecton’s founding team previously built the Michelangelo feature store at Uber. Willem Pienaar, who now serves as tech lead at Tecton, pioneered the development of Feast, one of the first open source feature stores. Provectus has deep expertise designing and building feature stores for various enterprise clients.

“Strategy-wise, composability, cloud agnosticism, and a multi-cloud approach are fundamental guidelines for the CIOs of modern enterprises. The fact is, machine learning infrastructure is not an exception,” says Pushkarev. “For Provectus, collaboration with Tecton and Feast is a chance to unite the community towards a better machine learning stack for enterprises.”

In the past few years, feature stores have evolved considerably; Provectus, Feast, and Tecton share a common vision of their future, which is outlined in the recent blog posts – “A State of Feast” and “Feature Store as a Foundation for ML”. A new version of Feast has recently emerged and was announced by Willem Pienaar & Jay Parthasarthy. At the apply() conference, the Provectus team gave a presentation on the roadmap for Feast and AWS integration.

About Provectus
Provectus is an Artificial Intelligence consultancy and solutions provider, helping companies in Healthcare & Life Sciences, Retail & CPG, Media & Entertainment, Manufacturing, and Internet businesses achieve their objectives through AI. Provectus is headquartered in Palo Alto, CA. For more information, visit provectus.com.

About Tecton
Tecton provides an enterprise-ready feature store for machine learning that enables organizations to manage the complete lifecycle of features, from engineering new features to serving them in production for real-time predictions. Tecton is headquartered in San Francisco, CA. For more information, visit tecton.ai.

About Feast
Feast is an open source feature store for machine learning that helps data scientists and ML engineers bridge the gap between data and machine learning models. It provides the fastest path to production for ML features. For more information, visit feast.dev.

Related