Databricks has expanded its Mosaic AI platform to help enterprises build and deploy large language models (LLMs). The expansion follows Databricks' acquisition of MosaicML for $1.3 billion a year ago. The rebranded Mosaic AI now includes tools for fine-tuning and pre-training models, building custom LLMs with enterprise data, and evaluating compound AI systems. Databricks' VP of Engineering, Patrick Wendell, highlighted these developments during the Data AI Summit, where the company also introduced a new AI tool catalog integrated into Unity Catalog. Additionally, the new Mosaic AI Agent Framework allows for the creation of agentic and RAG applications. The enhancements aim to empower businesses to develop robust AI systems and applications.