
Anatomy of a Tool-Calling Agent Loop: What's Happening Under the Hood
A Framework-Free Walkthrough of the Control Loop Behind Every Tool-Calling AI Agent
12+ years building production systems — most recently AWS serverless architecture for a large-scale data engineering platform. Currently extending that infrastructure work into agentic AI: building and deploying LLM-based agents, not just prototyping them. This blog covers what I build, what breaks, and the decisions behind both — written from actual deployment work, not tutorials.