🎥 Recorded live at the MLOps World | GenAI Summit 2025 — Austin, TX (October 8, 2025) Session Title: RAG Architecture at Capital One Speaker: Vaibhav Misra, Director & Distinguished Engineer, Capital One Abstract: Retrieval-Augmented Generation (RAG) has become a cornerstone for enterprise AI systems — but building it right requires more than connecting an LLM to a vector database. In this lightning talk, Vaibhav Misra from Capital One breaks down how his team designed and deployed a robust RAG architecture that enhances reliability, efficiency, and domain-specific accuracy in production. He shares practical lessons on overcoming the shortcomings of LLMs, structuring RAG data pipelines with vector search, and combining prompt engineering with fine-tuning to improve performance. What you’ll learn: • The key limitations of LLMs and how RAG helps overcome them • How to design a scalable, production-ready RAG pipeline • Practical steps for integrating vector search and fine-tuning • Strategies for improving retrieval accuracy and model reliability.