A year after the acquisition of MosaicML for $1.3 billion, Databricks is rebranding the platform as Mosaic AI and integrating it deeply into its AI solutions. Today, at the Data + AI Summit, the company is announcing the launch of several new features aimed at addressing key concerns of its customers.
Databricks co-founders CEO Ali Ghodsi and CTO Matei Zaharia discussed the significance of these developments ahead of the official announcements. Ghodsi emphasized the continued focus on improving model quality, cost-efficiency, and data privacy, which remain top priorities for users. The new Mosaic AI tools aim to address these concerns comprehensively.
The five new Mosaic AI tools being launched at the conference are:
Mosaic AI Agent Framework
Mosaic AI Agent Evaluation
Mosaic AI Tools Catalog
Mosaic AI Model Training
Mosaic AI Gateway
These tools are designed to streamline AI deployment processes, enhance model reliability, optimize costs, and ensure data privacy for Databricks' customers. Zaharia highlighted the importance of modular systems in deploying large language models (LLMs) into production, emphasizing the need for multiple components and external tools to augment foundation models with proprietary data.
Databricks is introducing the Mosaic AI Agent Framework and the Mosaic AI Tools Catalog to facilitate the development and deployment of these compound systems. The AI Agent Framework leverages the company's serverless vector search functionality to empower developers in building retrieval-augmented generation (RAG)-based applications efficiently.
These announcements mark a significant step forward in Databricks' mission to empower enterprises with high-impact, mission-critical AI applications. By providing comprehensive tools and services, Databricks aims to enable developers to build modular AI systems that are reliable, cost-effective, and privacy-preserving, setting the stage for the future of AI innovation.
