Course Objective & Scope
Loading a single agent with too many tools and responsibilities lowers quality, yet unnecessary multi-agent architectures increase cost and the surface for errors. This course addresses, at the architecture level, when more than one agent is needed and how those agents should be coordinated.
Supervisor, hierarchical teams, orchestrator-worker and swarm patterns, context transfer between agents, parallel subtasks and result aggregation are implemented with LangGraph, CrewAI and Microsoft Agent Framework examples. Participants build a multi-agent flow for an organisational scenario and measure it against a single-agent alternative.
Who Should Attend?
- AI engineers and architects designing complex agent systems
- Platform teams building an agent platform
- Technical leads deciding on an agent framework
- Prerequisite: experience with tool calling and at least one agent framework
Key Takeaways
- Justify when a multi-agent architecture is needed
- Implement supervisor, hierarchical and swarm patterns
- Design context transfer and isolation between agents
- Aggregate results from parallel subtasks reliably
- Measure the cost, latency and quality of a multi-agent system
Curriculum
Single Agent or Multiple Agents?
- Tool count, specialisation and context window limits
- Cost and failure risks of multi-agent systems
Orchestration Patterns
- Supervisor and hierarchical teams
- Orchestrator-worker and parallel subtasks
- Swarm and network topologies
Context and Communication
- Shared state versus message-based communication
- Context isolation and transfer through summarisation
Framework Comparison
- LangGraph, CrewAI and Microsoft Agent Framework approaches
- Selection criteria and interoperability
Measurement and Organisational Application
- Per-agent traces, cost and success metrics
- A multi-agent flow for an organisational scenario compared with a single agent
Other Courses in This Category
-
Hands-on training on designing working, episodic, semantic and procedural memory layers for AI agents using context summarisation, vector and graph memory stores, memory consolidation and retention policies.
- Live Course
- 5 modules
-
Technical course on safely executing LLM- and agent-generated code using gVisor, Firecracker microVMs, container isolation, network/filesystem restrictions and resource quotas.
- Live Course
- 5 modules
-
Course on evaluating end-to-end task completion, trajectory quality and tool selection in AI agents, and setting up distributed tracing and live monitoring with OpenTelemetry, LangSmith, Arize Phoenix and Langfuse.
- Live Course
- 5 modules
-
Training on protecting multi-step agents against API outages, rate limits, partial failures and crashes using retries, model fallbacks, idempotency, compensating actions and durable execution with Temporal and LangGraph.
- Live Course
- 5 modules
-
Architectural design course examining prompt chaining, routing, parallelisation, orchestrator-workers, evaluator-optimizer and autonomous agent patterns, and how to pick the simplest effective pattern for each problem.
- Live Course
- 5 modules
-
Hands-on developer training on combining ReAct reasoning loops, Plan-and-Execute task planning and explicit state/checkpoint management into a single unified hybrid agent architecture.
- Live Course
- 5 modules
Get in touch
