Learn how to optimise your RAG performance. The session covers 12 methods and best practices like semantic search and query expansion to improve performance of your LLM-based apps, personalize chatbot interactions, and reduce hallucinations. You’ll learn how to: 1. Optimise retrieval quality of your RAG setup 2. Use follow-up questions to refine retrieval dynamically based on user question 3. Integrate conversation memory to personalise responses and many more! Hosted by: ‪- Lena Shakurova‬, Founder & CEO of ParsLabs - Claire Longo, AI Researcher at @comet_opik, Mathematician and AI Researcher with over a decade of experience building AI models and leading engineering teams across enterprise companies and startups. She is currently a Developer Advocate at Comet. Claire holds a Bachelor’s in Applied Mathematics and a Master’s in Statistics from the University of New Mexico. Beyond her technical work, she is a speaker, advisor, and podcast host, dedicated to mentoring engineers and data scientists while championing diversity and inclusion in AI. Timestamps 00:00 Intro 02:14 Intent + RAG as fallback 04:18 Function calling 07:25 Use FAQs instead of raw data 11:16 Two stage similarity search 12:33 Query rephrasing 13:46 Multi-query retrieval or query expansion 15:35 Chat history summary 17:35 RAG on past conversation history 20:37 Fluctuating & stable properties 25:35 Clarifying questions 27:55 Outro 🤝 Need expert help with your current setup? Schedule a consultation with me: ☕ Support the work I do: