Databricks today announced several innovations to its Mosaic AI platform to help customers build production-quality generative AI applications. The company is investing in three key areas: support for building complex AI systems, capabilities to improve model quality, and new AI governance tools.
Organizations struggle to move their generative AI projects from pilot to full-scale production due to privacy, quality, and cost concerns. While the underlying models have improved significantly, they still face the challenge of producing consistently high-quality results. To address these issues, organizations are deploying composite AI systems rather than just deploying a single large-scale model.
This approach leverages multiple components, including a range of models, retrievers, vector databases, and tools for evaluation, monitoring, security, and governance, to improve production quality and enable more accurate, safe, and governed AI applications.
“We believe that complex AI systems are the best way to maximize the quality, reliability, and measurement of AI applications going forward, and could be one of the most important trends in AI in 2024,” said Matei Zaharia, co-founder and CTO at Databricks. “Databricks is uniquely positioned to capitalize on these trends by investing in improving quality and augmenting models with real-time data, agents, and tools to provide new, little-known capabilities to models.”
Databricks is releasing several new features to help customers build production-quality generative AI applications.
Mosaic AI Agent Framework and Mosaic AI Tools Catalog Helps organizations build composite AI systems. The agent framework enables developers to quickly and securely build high-quality Retrieval-Augmented Generation (RAG) applications using foundational models and enterprise data. The tool catalog enables organizations to manage, share, and register tools using the Databricks Unity Catalog, ensuring safe and governed use of tool-enabled models.
Mosaic AI Quality Lab is an AI-assisted evaluation tool that automatically determines if the output is of high quality and provides an intuitive UI for collecting feedback from human stakeholders, empowering organizations to deploy production-quality generative AI solutions.
Mosaic AI Model Training Open source foundational models can be fine-tuned with an organization's private data to deliver higher quality results for specific use cases. These fine-tuned models are fully owned and managed by the customer, and are faster and less expensive to deliver than large-scale proprietary models.
Mosaic AI Gateway It provides a unified interface to query, manage, and deploy open source or proprietary models, enabling customers to easily switch between Large Language Models (LLMs) powering their applications without complex code changes, and provides usage tracking, guardrails, governance, and monitoring to ensure quality and control spend.
Several Databricks customers, including Corning, Ford Direct, and Lippert, are already benefiting from these new capabilities in building generative AI applications, leveraging the Databricks Data Intelligence Platform and Mosaic AI to improve ingest speed, response quality, accuracy, and reliability in deploying to production environments.
The new Mosaic AI capabilities are part of Databricks' ongoing commitment to help customers harness the power of generative AI while maintaining data privacy, quality, and cost-efficiency. As organizations continue to explore the possibilities of AI, Databricks aims to provide them with the tools and platform they need to build enterprise-ready AI applications.
