Agents and MCP for PostgreSQL

​AI agents need more than just prompts - they work with tools and data. This presentation explores the evolution of agentic workflows and the specific mechanisms they use to interact with databases. We’ll focus on the Model Context Protocol (MCP) and how it connects AI to PostgreSQL, moving beyond basic RAG to more advanced data driven tools.

​While RAG (Retrieval-Augmented Generation) is well-understood, it is often a one-way street. The next frontier is agents allowing models to interact with data, query, and reason over databases dynamically. The Model Context Protocol (MCP) is an industry standard that solves the fragmentation of tool-calling. This talk provides a practical, "how-to" guide for the popular open-source tools in that field and how they interact with PostgreSQL.

​Speaker: Gleb Otochkin

​Gleb, a Cloud Advocate at Google, specializes in database technologies and AI integration within data-driven applications. His expertise includes relational databases, data transformation and ETL, application development, data replication and integration solutions. He is a runner completing 2-4 marathons per year. In the past he participated in expeditions dedicated to oceanic research in the Pacific as oceanologist.