https://www.infoq.com InfoQ Retrieval-Augmented Generation Presentations feed https://www.infoq.com/presentations/rag-vllm/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=Retrieval-Augmented+Generation-presentations <img src="https://res.infoq.com/presentations/rag-vllm/en/mediumimage/stephen-batifol-medium-1758016495689.jpg"/><p>Stephen Batifol discusses building and optimizing self-hosted, multimodal RAG systems. He breaks down vector search, nearest neighbor indexes (FLAT, IVF, HNSW), and the critical role of choosing the right embedding model. He then explains vLLM inference optimization (paged attention, quantization) and uses Mistral's Pixtral to detail multimodal large language model architecture.</p> <i>By Stephen Batifol</i> QCon London 2025 Transcripts Large language models Retrieval-Augmented Generation AI, ML & Data Engineering presentation Fri, 10 Oct 2025 14:12:00 GMT https://www.infoq.com/presentations/rag-vllm/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=Retrieval-Augmented+Generation-presentations Stephen Batifol 2025-10-10T14:12:00Z /presentations/rag-vllm/en