AutoGen LLM Agents Training in Bangkok, Thailand

The AutoGen LLM Agents Training is an advanced, hands-on program designed to teach participants how to build multi-agent systems using Microsoft AutoGen and similar LLM agent frameworks.

The course focuses on designing, orchestrating, and deploying collaborating AI agents for complex tasks such as research, coding, workflow automation, and decision-support systems.

Introduction to Multi-Agent LLM Systems

• What are LLM agents?
• Single-agent vs multi-agent systems
• Benefits and challenges of multi-agent architectures
• Use cases for collaborating agents

Overview of AutoGen Framework

• What is Microsoft AutoGen?
• Core components and abstractions
• Agent types and conversation patterns
• Typical AutoGen architecture

Environment Setup & Project Structure

• Installing AutoGen
• Project setup and dependencies
• API key management
• Development environment best practices

Building Your First AutoGen Agents

• Creating basic agents
• Defining agent roles and system messages
• Agent-to-agent communication
• Running simple multi-agent conversations

Agent Roles, Personas & Responsibilities

• Designing specialized agents (planner, coder, reviewer, etc.)
• Role-based prompting strategies
• Responsibility separation
• Reducing agent overlap and confusion

Tool & Function Integration

• Connecting agents to tools
• File system and code execution tools
• API and service integrations
• Safe tool execution

Multi-Step & Multi-Agent Workflows

• Task decomposition across agents
• Orchestrating agent collaboration
• Managing dependencies and handoffs
• Workflow state management

Memory & Context Sharing

• Shared vs private agent memory
• Context passing between agents
• Long-running conversations
• Preventing context overload

Human-in-the-Loop & Supervision

• Approval and review checkpoints
• Interrupting and correcting agents
• Supervisory agents
• Reducing risk in autonomous systems

Error Handling & Debugging Agents

• Common agent failure modes
• Debugging multi-agent conversations
• Recovering from tool failures
• Improving agent reliability

RAG & Knowledge Integration

• Connecting agents to knowledge bases
• Retrieval-Augmented Generation for agents
• Document and database tools
• Grounding agent responses

Testing & Validation of Agent Systems

• Unit testing agent logic
• Regression testing conversations
• Evaluating agent performance
• Quality assurance for agent workflows

Security, Privacy & Governance

• Securing tool access
• Handling sensitive data
• Audit logs for agent actions
• Governance models for agent systems

Performance, Cost & Scaling

• Scaling multi-agent systems
• Parallel agent execution
• Token and cost optimization
• Latency management

Production Deployment Patterns

• Deploying AutoGen systems
• CI/CD for agent-based apps
• Monitoring and observability
• Versioning agents and workflows

Hands-on Exercises

Summary and Conclusion