May 23
Loading video player...
This is Video 2 of my AI Engineering series. In Video 1 we gave an LLM the ability to call a single Python function. In this video we go one level deeper - the function itself is an entire AI agent. We build a multi-agent system from scratch using nothing but the raw OpenAI SDK and a while loop. An orchestrator agent delegates to two specialist sub-agents - a researcher and a critic - to write and refine a blog introduction. This video is great if you’re trying to understand how modern AI agent systems actually work under the hood. What you will learn: - How to wrap an LLM call inside a plain Python function to create a sub-agent - How to write tool schemas for your sub-agents - How the orchestrator loop works when multiple tools are available - How globals() dynamically calls functions by name - Why one agent calling another is the foundation of every multi-agent system ever built GitHub Repo: https://github.com/Prashanna-Raj-Pandit/Agentic-AI-YouTube/blob/main/openai%20function%20calling/agent_as_a_tool.py LinkedIn: https://www.linkedin.com/in/ra-prashanna/ #AI #OpenAI #Python #AIAgents #MachineLearning #GenerativeAI #AIEngineering #BuildingInPublic