orchestration patterns

Post explicitly identifies orchestration as one of five core design components in Compound AI Systems

Introduces orchestrator-worker pattern for multi-agent coordination as architectural approach to managing context flow across specialized agents.

Supervisor-subagent pattern is a specific multi-agent orchestration architecture with context management implications.

Reference architecture (Router → Orchestrator → Specialists → Synthesizer → Checker) is specific instantiation of orchestration pattern with explicit context management layers

The six patterns (sequential, parallel, hierarchical, etc.) are orchestration approaches that directly affect how context flows between agents

The three-layer orchestration model (orchestrator → agents → coordination layer) is a fundamental pattern for how to structure agent systems to manage context flow and task decomposition.

Subagent architecture is a specific orchestration pattern where context flows hierarchically through a central orchestrator. The analysis of its overhead/control tradeoff extends understanding of orch

The orchestration section describes control flow, routing, workflows, and message queues—a practical taxonomy of orchestration concerns for multi-agent systems.

LangGraph's graph-based node/edge model is a specific orchestration pattern for managing agent sequencing and information flow.

The hybrid approach (CrewAI task nested in LangGraph node) is a multi-level orchestration pattern. Shows composability of abstractions.

LangGraph's graph-based architecture is an orchestration pattern for coordinating how agents process and act on context; relates to multi-agent coordination