Building LLM Agent Systems with AutoGen Training Course

Foundations of LLM Agent Systems

• Concepts of LLM agents and multi-agent architecture
• Overview of the AutoGen framework and its ecosystem
• Core agent roles: user proxy, assistant, function caller, and others

Installation and AutoGen Configuration

• Setting up the Python environment and required dependencies
• Essentials of AutoGen configuration files
• Establishing connections to LLM providers (OpenAI, Azure, local models)

Agent Architecture and Role Definition

• Exploring agent types and conversational patterns
• Defining agent objectives, prompts, and operational instructions
• Implementing role-based task delegation and control flow

Function Calling and Tool Integration

• Registering functions for agent utilisation
• Executing functions autonomously and collaboratively
• Integrating external APIs and Python scripts with agents

Conversation Control and Memory Management

• Tracking sessions and maintaining persistent memory
• Facilitating agent-to-agent messaging and token processing
• Governing conversation context and history

Comprehensive Agent Workflows

• Creating multi-step collaborative tasks (e.g., document analysis, code review)
• Simulating user-agent dialogues and decision-making chains
• Debugging and optimising agent performance

Applications and Deployment Strategies

• Internal automation agents: research, reporting, scripting
• External-facing bots: chat assistants, voice integrations
• Packaging and deploying agent systems for production environments

Wrap-up and Further Development