Build persistent and scalable AI agent memory with TiDB | ODSP918

This session demonstrates how to build persistent, scalable memory infrastructure for AI agents using TiDB, which combines vector search, BM25 full-text search, and SQL in a unified database table. Targeted at developers building agentic AI systems, it covers hybrid retrieval techniques, ACID-compliant transactions for multi-table operations, and deployment on Azure with Azure OpenAI embeddings. Includes a case study of Manus AI deploying millions of agent databases.