Today marks a big milestone for Tecton. We’re excited to announce that the Tecton feature store is now in General Availability, and that we have raised $35 Million in Series B funding from our existing lead investors, Andreessen Horowitz and Sequoia, to fuel our next phase of growth.

We started Tecton in late 2018 with the vision to make it easy and safe to build smart product experiences with machine learning. This means bringing the best practices of machine learning development to everyone, and enabling every team to build with the speed, trust, and power of the industry’s most capable applied ML organizations.

The Tecton founding team met at Uber while creating the Michelangelo platform. At Uber, we experienced first-hand that the hardest part of getting ML to production is the data. Data scientists spend the majority of their time cleaning data and engineering features. They typically develop in silos using tools optimized for experimentation, but not for production, often relying on separate data engineering teams to implement and maintain production-grade versions of the feature pipelines. This re-implementation process can painfully add weeks or months to ML efforts, especially in cases requiring streaming or real-time data. The extremely high cost of operationalizing and maintaining ML feature pipelines means ML teams’ best models often don’t reach production.

On the Michelangelo team, we developed the feature store to solve these problems. It quickly became an essential component of the stack for teams building operational ML applications, delivering operational excellence, enforcing compliance, and enabling collaboration. However, until recently, features stores have only been available to technology giants with very large ML infrastructure teams. We set out to change that, and make feature stores available to every organization.

Over the past two years, we’ve been very busy building an enterprise-grade feature store with input from early access customers. Our vision has always been to bring the principles of DevOps to ML data and allow data scientists to own their features from development to production. With that objective in mind, Tecton is now an enterprise-grade feature store that combines:

  • A minimal and simple SDK for configuring feature transformation and serving pipelines using popular data science libraries and modern batch, streaming, and real-time data processing infrastructure
  • End-to-end management and orchestration of the complete feature lifecycle, from feature creation and evaluation, to deployment, sharing, and monitoring
  • Built-in best practices for feature pipeline operationalization (logging, auditing, lineage tracking, real-time serving, drift monitoring, etc.)
  • Built-in software engineering best practices and integrations with existing MLOps and DevOps tooling
  • Governance workflows giving organizations control of how features are shared, used, and reused
  • Enterprise-grade security and scalability, supporting tens of thousands of features and millions of predictions per second
  • Delivery as a cloud-native service that is fully-managed by Tecton, elastically scalable, and cost-optimized for the cloud

Along the way, we’ve had the opportunity to work with a great set of customers including Atlassian, Tide, Branch Metrics, and many others  ranging from large Fortune 50 companies to small startups. Many of these teams are now running Tecton in production to support mission-critical applications. Thank you for providing invaluable feedback and helping guide Tecton’s roadmap so far. Tecton is now an enterprise-grade, production-ready feature store available on AWS and we’ll soon be available on Azure and Google Cloud too! 

We’re definitely not done yet. The MLOps landscape has changed extensively over the past 12 months. In that time, the demand for features stores has risen dramatically. We’ve been listening to your requests and have big plans and a packed roadmap for the next year. 

We’re grateful to our existing lead investors, Martin Casado of Andreessen Horowitz and Matt Miller of Sequoia, for the confidence they’ve placed in Tecton and for supporting us along the way. After leading our Series A round earlier this year, they’ve now led Tecton’s Series B fundraise which provides a fresh infusion of $35 Million of capital. Our Series B gives us the resources to build the best, uncompromising feature store to allow every data science team to operationalize ML at scale. The capital will be used to accelerate our roadmap, which includes bringing Tecton to all the major clouds, supporting a broader range of data processing frameworks, and simplifying the feature engineering experience for data scientists.

We’re incredibly excited for what 2021 will bring for MLOps and feature stores. If you’d like to take Tecton for a spin, don’t hesitate to request a free trial and we’ll get you onboarded in no time. If you’d like to learn more about Tecton, schedule a demo and we’ll be in touch. Or, if you’re interested in helping us build the future of data for ML, check out our job openings – we’re hiring across all our teams.