Build and operate AI-native backend services that orchestrate multi-step, stateful LLM agents for GTM workflows. Design model-agnostic abstractions, integrate external systems, deploy containerized services on Kubernetes, ensure observability, and run experiments to measure effectiveness.