AI Agents in Python: From Basics to Multi-Agent Systems
| Speaker | Dirk Brand |
|---|---|
| Track | Applications of LLMs and AI |
| Type | Regular talk (45 minutes) |
Abstract
This talk provides a practical introduction to building AI agents in Python, using the Agno framework for building intelligent, autonomous agents. We will progress systematically through 5 levels of agent sophistication, starting with foundational concepts and building toward advanced multi-agent systems. We'll begin with Level 1 agents equipped with tools and instructions, then advance through agents enhanced with knowledge bases and persistent storage (Level 2), followed by agents capable of memory retention and sophisticated reasoning (Level 3). Each concept will be demonstrated through coding examples in the Agno environment, showing real implementations and practical patterns that attendees can immediately apply.
Building on individual agent capabilities, we'll explore Level 4 Agent Teams that can reason collectively and collaborate within sophisticated multi-agent architectures. These systems showcase how individual agents can work together using coordination patterns like routing, collaboration, and orchestration to tackle problems beyond the scope of single agents. We will then cover Level 5 Agentic Workflows with state management and deterministic execution, demonstrating how to orchestrate complex multi-agent systems for real-world problem-solving scenarios including distributed task processing, resource allocation, and collaborative decision-making across diverse domains.
The presentation concludes with production-ready best practices for deploying agentic systems at scale, including monitoring strategies, performance optimization, and architectural considerations for enterprise environments. This talk will be most valuable to AI enthusiasts and AI developers seeking to understand the current landscape of agentic software and the practical capabilities of these systems. You will gain a comprehensive understanding of what's possible with modern agent frameworks, practical implementation knowledge, and inspiration for future products that leverage the power of autonomous AI agents.
