A solo founder runs a five-continent tender platform on AlloyDB and MCP (cloud.google.com)

Lucius AI, a tender-intelligence startup, operates a comprehensive platform spanning five continents, utilizing AlloyDB for PostgreSQL to manage a database that consolidates relational catalogs, audit logs, and vector embeddings into a single managed system. This integration has significantly streamlined operations for founder Davor Jerković, allowing him to cover over 210,000 international tenders with minimal overhead. Notably, by migrating semantic search to a ScaNN index, query latency was reduced from 1.14 seconds to just 24 milliseconds, marking a remarkable 47x speed improvement and enabling quicker data retrieval for users bidding on public contracts. The use of the Model Context Protocol (MCP) further enhances operational efficiency by automating routine database tasks such as performance optimization, incident forensics, and database quality checks, while maintaining strict access controls. This innovative blend of tools not only alleviates the need for a dedicated data engineering team but also demonstrates how AI can empower a single operator to successfully manage a complex, global platform. Looking forward, Lucius AI plans to enhance its architecture further with automated vector embeddings and a fully managed MCP server, continuing to push the boundaries of what is possible with AI in operational management.