Orchestrating Large Language Models (LLMs) with CrewAI
summary
Crew AI is a Python framework that wires agents, tasks, tools and crews so LLMs become orchestrated workers rather than one-off prompts; it is LLM-agnostic and exposes tools (serp/search, custom API hooks), enforces structured outputs via pydantic, and adds guardrails and conditional or hierarchical orchestration to reduce nondeterminism. Examples shown: a ski-trip optimizer that enforces budget constraints with a manager-agent and it runs repeatedly, and a Streamlit AI mock-interviewer using OpenAI and Whisper that works end-to-end.
AI-written summary · about
chapters (10)
- 0:00Introduction and setup
- 3:06History of LLMs and agents
- 8:43What are AI agents
- 13:03Why CrewAI for orchestration
- 16:38Creating agents and tasks
- 21:19Orchestration types in CrewAI
- 23:05Tools and guardrails
- 28:22Ski trip use case
- 37:11Live demo of AI mock interviewer
- 43:09Q&A session
related talks
01h 04mAgentic AI: Build a Multi-Agent Application with CrewAI
29mDemystifying AI Agents with Python Code
36mLLMs and AI agents demystified
41mBringing stories to life with AI, data streaming and generative agents
01h 04mBuilding Expertise Into Generated Conversations
32mEvent-Driven & Multi-Agent Architectures for Gen AI Workflows
11mLLM Agents as Co Pilots for Scalable Decision Systems
01h 43mBuilding LLM Agents (Hackathon)