DataStax has introduced new tooling to help customers build GenAI apps on its data platform, including the open source Cassandra database. The company acquired Langflow, a visual framework for building AI applications, and is previewing it in its Astra Cloud platform. DataStax also offers integrations with GenAI tools like OpenAI and Hugging Face, as well as a system called RAGStack 1.0 for enterprise-scale implementation of retrieval-augmented generation.

DataStax’s Chief Marketing Officer, Jason McClelland, highlighted the challenges organizations face in building AI applications due to rapidly iterating tooling. DataStax aims to provide integrated tooling for data preparation, vector embedding, data ingestion, and data chunking. However, other database vendors like Couchbase, MongoDB, and Oracle are also offering tooling for AI application development.

The question for customers using multiple databases is which vendor to choose as the main support for GenAI application development. IDC research vice president Carl Olofson noted that DataStax’s tooling makes RAG easier and faster, but it raises the question of whether it enhances the use of Cassandra or offers an alternative sales track.

Tech advisory and research company ISG’s director, Matt Aslett, emphasized that enterprises will continue to use multiple databases for different applications and use cases, even with the rise of generative AI. While starting a pilot project with an existing data platform may make sense, organizations investing heavily in AI may have separate teams for AI-specific platforms and tooling, independent of a specific database vendor.

Source link

Source link: https://www.theregister.com/AMP/2024/07/01/database_ai_training/