May 15
Loading video player...
In this video you will build a hierarchical multi-agent system using CrewAI where an autonomous manager LLM orchestrates a team of specialised agents across different domains. This is Video 6 of the Building AI Agents with CrewAI series. Up until now every crew in this series has used Process.sequential, where tasks execute one after another in a fixed order. In this video you switch to Process.hierarchical, where a manager LLM is responsible for delegating tasks, coordinating agent execution, and synthesising results across the entire crew. The example builds a startup planning crew with three specialised agents: a Marketing Strategist, a Tech Lead, and a Finance Analyst, each handling a separate domain. The manager LLM oversees the full workflow without you hardcoding the execution order. This is how real organisational workflows are modelled in agentic systems. —————————————————————————— 🧠 What you'll learn —————————————————————————— The difference between Process.sequential and Process.hierarchical How the manager_llm parameter enables autonomous task delegation How to design domain-scoped agents that operate independently How hierarchical orchestration scales to complex multi-agent workflows —————————————————————————— 📦 Stack —————————————————————————— CrewAI Gemini 2.5 Flash (via LiteLLM) python-dotenv —————————————————————————— 🔗 Links —————————————————————————— CrewAI Docs: https://docs.crewai.com Video 1: https://youtu.be/brexE2qNQXo Video 2: https://youtu.be/2kXhBxmjA_Y Video 3: https://youtu.be/WNmz0qUJ4UM Video 4: https://youtu.be/ZOgMPEdDrcw Video 5: https://youtu.be/T9vezBAyAjg #CrewAI #AIAgents #Gemini #Python #LLM #AIEngineering #AyonaireAcademy #BuildInPublic #GenerativeAI