Is Graph Engineering just reinventing systems architecture for the AI age?

A recent article on Dev.to explores the emerging concept of 'graph engineering' in the context of AI development. Unlike knowledge graphs or agent swarms, graph engineering focuses on applying traditional backend architecture principles—such as microservices, strict schemas, and fan-out/join patterns—to multi-agent systems. By moving away from unpredictable 'black-box' loops, developers can use frameworks like Google's Agent Development Kit (ADK) to create more reliable, debuggable, and structured AI workflows. The piece features an interview with Annie Wang, who explains how these graph-based patterns provide the predictability required for enterprise-grade AI applications. The author argues that this approach essentially brings the rigor of software engineering to the chaotic nature of LLM-based agents, allowing for better control over shared state and complex automated pipelines. Readers are encouraged to explore the provided ADK codelab to see these architectural concepts in action.