Bringing AI to Real-Time Data. Interview with Stephane Castellani

Discover how CrateDB is connecting real-time databases with AI, from voice-driven text-to-SQL use cases on the factory floor to new standards (MCP), Retrieval-Augmented Generation (RAG), Table-Augmented Generation (TAG), and the future of agentic AI. In this interview, Stephane Castellani (SVP Marketing, CrateDB) explains: - why deep technical domain knowledge matters in marketing, - how CrateDB delivers sub-second ingestion, indexing, and queries for fast-moving data, - how LLMs are used to translate voice or natural language into SQL for non-technical users, - what RAG, TAG and MCP mean in practice, - and how CrateDB sees the next 5 years as AI becomes an orchestration layer on top of core data platforms. 0:00 Intro 0:15 Guest intro: Stephane Castellani (CrateDB) 1:00 Stephane’s technical background & move into marketing 5:00 Text-to-SQL, low-code & no-code discussion 6:40 Voice + LLMs on the factory floor (real-time use case) 7:43 What “real-time database” means and why it matters 8:19 CrateDB’s mission: sub-second ingest → query 9:24 RAG (Retrieval-Augmented Generation) explained 11:08 TAG (Table-Augmented Generation) & agentic AI foundation 12:21 MCP (method/connector protocol) — LLMs calling external services 14:00 Orchestration, arbitration & interpreting results for humans 17:46 Where CrateDB might be in 5 years — the future of data + AI 18:54 Humans + AI — productivity, governance and caution 21:23 Wrap up & thanks