How Skyscanner Runs Real-Time AI at Scale with Databricks

Explore how to deploy and scale AI models in production environments through this 32-minute conference talk featuring insights from Databricks and Skyscanner. Learn about the complexities of modern AI deployment, including different model types, tight timelines, and growing infrastructure demands. Discover how Mosaic AI Model Serving enables teams to efficiently deploy and scale both traditional machine learning and generative AI models with built-in monitoring and governance capabilities. Gain practical insights from Skyscanner's real-world experience integrating AI into their products, scaling to over 100 production endpoints, and developing the processes and team structures necessary to support AI at enterprise scale. Understand strategies for building compound AI systems that leverage models, feature stores, and vector search technologies. Master approaches to monitoring, debugging, and governing production AI workloads while maintaining low latency and minimal operational overhead. The session covers practical deployment techniques, scaling strategies, and organizational best practices for running AI systems in production environments.