April 2020 Summaries

The industry needs to solve DevOps for ML data, as most ML projects fail to deploy and operate in production. The main challenge lies in accessing the right raw data source, building features from it, combining features into training data, calculating and serving features in production, monitoring features in production, and addressing related issues like data leakage, time travel, freshness, and ownership. Tecton, a centralized data platform for machine learning, aims to fill this vacuum by providing a feature pipeline, feature store, feature server, SDK, web UI, and monitoring engine to help ML teams bring DevOps practices to ML data, ensuring planability, code quality, build reliability, testing, release management, deployment, operation, and monitoring.

Tecton is a data platform designed to make machine learning operational and accessible to every organization. It aims to bring best practices of machine learning development to teams worldwide, enabling them to build with the speed, trust, and power of industry-leading applied ML organizations. Tecton expands on the feature store architecture developed at Uber, managing the end-to-end lifecycle of features for ML systems that run in production. The platform helps teams develop standardized features, deploy and serve feature data safely, curate an organization-wide repository of vetted features, monitor feature data, and accomplish these tasks with best practices, reproducibility, lineage, logging, and trust needed for business-critical functions built on top of the system. Tecton has raised $25MM in funding and is hiring an outstanding engineering team to further develop its core technology and bring it to market.