Explore the complexities of building production-quality retrieval systems for AI applications in this 29-minute conference talk from the Linux Foundation's Open Source Summit. Delve into the anatomy of robust retrieval systems that go beyond simple vector search to create compelling, scalable AI solutions. Learn about various retrieval methods and their respective strengths and weaknesses, discover techniques for combining results from multiple search approaches, and understand how to generate features that enhance search performance. Gain insights into measuring and validating improvements in retrieval systems, ensuring each modification actually enhances performance. Master the key challenges AI teams face when implementing retrieval systems at scale, from basic vector search to sophisticated multi-method approaches that power production AI agents and applications.