Operationalizing AI at Scale with H2O Feature Store

AT&T and H2O.ai launched our co-developed H2O Feature Store in 2021 to act both as an AI water cooler facilitating discovery and sharing of relevant features across the enterprise and to streamline the productionalization of models through reusable and scalable data pipelines accelerating time-to-market from months to hours. It was a milestone in its own rite to have built an enterprise grade feature store for one of the largest corporations in the US, but that was not enough. Both organizations are mission driven to democratize AI and that has pushed the partnership to tackle the challenges of going from one large scale implementation to supporting maker organizations of all sizes as we productized the feature store making it an integral component to the H2O AI Cloud. This presentation will cover the origins and motivations for the feature store, the value AT&T has realized from having the feature store in production, key concepts and capabilities of the H2O Feature Store, feature stores as enablers and consumers of Generative AI, and our journey to 1.0 and beyond.