In this course we will get fair understanding of what AI agents are and how to use them to solve real world problems. The course covers the different types of AI agents that can be built. The different types including single, interactive, multi and autonomous agents. The course explains about multi tool agent systems and function/tool calling by the AI agents. As we progress we understand the architecture of AI agents.
We will also try to understand how to use different models from different providers to build a multi model agent using litellm. We also try to explore about building multi agent systems to avoid overloading one agent and distribute the load using multi agent systems. The course also explains about how to protect AI agents by adding sanitation logic using callbacks and also describes different types of callbacks. Use of callbacks for logging ,debugging, mentoring and guarding the AI system.
Use of Pydantic to handle data for communication between AI agents is explained to reduce cost and improve communication among Agents. The course slowly builds towards understanding the use of workflows and how workflows can be built to automate the processes. Explains the difference between graph based workflows and dynamic workflows. Dynamic workflows are used to build more complex workflows. The whole course revolves around the usage of Google-Adk and Claude-Sonnet and Gemini.
