As generative AI systems evolve from single LLM calls to complex, goal-driven workflows, multi-agent architectures are becoming important for building robust, scalable, and explainable AI applications.

This session presents a practical framework for designing and implementing multi-agent generative AI systems through four orchestration patterns that define how agents coordinate.

The Orchestrator–Worker pattern uses a central agent to break down a task and delegate subtasks to specialized worker agents, then aggregate and validate the results. The Hierarchical Agent pattern organizes agents into layers such as manager, specialist, and executor, which supports abstraction, delegation, and structured error handling. The Blackboard pattern allows agents to contribute to and react from a shared workspace, enabling loosely coupled, event-driven collaboration. The Market-Based pattern treats agents as autonomous participants that negotiate, bid, or compete for tasks and resources, which is useful in dynamic, resource-constrained environments.

For each pattern, the session explores practical use cases such as customer support triage, research synthesis, and code generation pipelines, and examines trade-offs in latency, complexity, and observability.

What You Will Learn

• The four orchestration patterns used to structure multi-agent AI systems
• How different coordination models support tasks such as customer support triage, research synthesis, and code generation pipelines
• Key trade-offs between patterns, including latency, system complexity, and observability

Who Should Attend

• Software Developers
• Software Architects
• AI and Machine Learning Engineers
• Technical Leads
• Engineering Managers