Feast

Feast is an end-to-end open source feature store for machine learning. It allows teams to define, manage, discover, and serve features.

Employees

1-10

Location

San Francisco, CA

AI Classification

DataAIData Infrastructure

Feast is an open-source feature store for machine learning and AI, maintained under the Linux Foundation, that enables engineering teams to define, manage, discover, and serve structured features for model training and real-time inference. It is widely adopted by enterprise ML teams at companies including Adyen, Shopify, NVIDIA, Robinhood, Walmart, and IBM.

Deep Intelligence

Buying Signals

Red Hat integrated Feast into OpenShift AI 2.20 as a technology preview (November 2024), signaling accelerating enterprise platform adoption and ecosystem legitimacy
Feast officially joined the PyTorch Ecosystem, expanding integration with the dominant ML framework and broadening the addressable developer base
Oracle Database became a fully featured Feast offline store with Kubernetes-native operator support, extending reach into Oracle-heavy enterprise environments
Feast Operator enhanced with HPA autoscaling, KEDA, PodDisruptionBudgets, and topology spread constraints — signals active maturation for production-grade deployments
12M+ downloads and 5,500 active Slack community members indicate sustained and growing real-world adoption across organizations

Account Scoring

Tier BLow Churn RiskProduct Launch

Pain Signals

As a free open-source project, Feast provides no enterprise SLAs, managed hosting, or dedicated support — teams with production uptime requirements must self-operate or migrate to commercial alternatives like Tecton
No G2 profile or commercial marketplace listing found, limiting discoverability for procurement teams running formal software evaluations
Tecton — a direct commercial competitor — was co-founded by the same alumni who created Feast's inspiration (Uber's Michelangelo), creating a recurring narrative that enterprise teams may 'graduate' from Feast to Tecton for managed reliability