Learn more about the path we’ve taken to shape the future of machine learning.
Our founding team first met at Uber, where they built the Michelangelo ML platform. Michelangelo was instrumental in enabling Uber to scale to 1000s of models in production in just a few years, supporting a broad range of use cases from real-time pricing, to fraud detection, and ETA forecasting.
At the heart of Michelangelo was the industry’s first feature store, designed to enable data teams to build ML features and serve them in production quickly and reliably.
Unfortunately, feature stores were only accessible to technology giants like Uber, Google, Facebook, and Amazon because they could leverage access to the best engineering talent available. The majority of organizations without the manpower to build feature stores still relied on armies of data engineers to monitor and run feature pipelines. Not only did they spend countless weeks — if not months — building and maintaining disparate data pipelines, but many Operational ML solutions never reached production environments at all.
Driven by the key insight that data was the hardest part of getting models to production, we founded Tecton in 2019 to democratize this technology and make feature stores accessible to every organization.
Our first commercial feature store was first unveiled in 2020. Since then, Tecton has become an award-winning feature platform that has partnerships with industry-leading companies like Snowflake and Databricks. Dozens of organizations depend on us to build reliable and powerful new features that accelerate, organize, and manage their path to production at scale.
