Python for Autonomous Agents: Tooling, SDKs & Orchestration Training Course

Introduction to Python Environments for Agentic Development

• Setting up Python, virtual environments, and dependency management
• Utilising Git and Docker for version control and isolation
• Best practices for maintaining reproducible environments

Overview of Agent SDKs and Frameworks

• LangChain, AutoGen, and other emerging SDKs
• Agent structure and lifecycle: perception, reasoning, and action
• Comparing SDK capabilities and architectural styles

Building Functional Agents in Python

• Creating a simple agent with LangChain
• Connecting agents to external tools and APIs
• Managing input/output, memory, and persistence

Tool and API Integration

• Defining and registering tools for agent use
• Secure API integration and key management
• Leveraging external data sources and custom function calls

Agent Orchestration and Communication Patterns

• Multi-agent collaboration using AutoGen
• Task delegation and planning logic
• Event-driven and asynchronous orchestration

Testing, Debugging, and Observability

• Testing agents with mock inputs and controlled environments
• Debugging message flow and tool invocation
• Implementing structured logging and performance metrics

Deployment and Production Considerations

• Packaging and containerising Python agent services
• Integrating with CI/CD pipelines
• Scaling, monitoring, and maintaining long-running agents

Summary and Next Steps