Build 3 PRODUCTION AI Agents in Python - Full Course (Agentspan)

This comprehensive course teaches viewers how to architect and build three different types of AI agents—a simple conversational agent, a RAG-based agent with company database access, and a complex multi-agent orchestrator—using the open-source Agentspan framework. The tutorial covers key architectural concepts such as durable execution, human-in-the-loop workflows, and real-time observability, ensuring that the agents built are stable and scalable for production environments.

Following a step-by-step approach, the instructor walks through installation, configuration, and implementation, specifically emphasizing the separation between the 'worker' (local execution logic) and the 'server' (state and orchestration management). Viewers learn how to incorporate tools, implement memory, add guardrails to prevent issues like prompt injection, and create structured, predictable outputs using Pydantic models.

The final segments move into advanced topics, demonstrating how to scale agents by deploying them with PostgreSQL and Docker Compose. The tutorial concludes with a practical guide on testing agent workflows using simulated events to ensure durability and resilience without relying on frequent, expensive LLM API calls, providing a complete pipeline for professional AI development.