Agentic AI

Build a Multi-Agent Application with CrewAI

Developed by Alessandro Romano, this workshop represents the culmination of hands-on training sessions delivered at PyCon DE, Data Hack Summit, and The Neural Maze. This platform provides a comprehensive resource for the workshop, featuring detailed explanations of all assignments and their prerequisites. The workshop content is based on this presentation, which provides the theoretical foundation for the practical exercises you'll find here.

Setup Prerequisites

OpenAI API Key

Required for accessing GPT models

Get API Key

SerperDev API Key

Required for web search capabilities

Get API Key
Development Environment Options

While I recommend using uv for the best experience, you're free to use any Python environment you prefer, such as Jupyter Notebook, Google Colab, or your existing setup.

1

Install uv (if not already installed)

curl -LsSf https://astral.sh/uv/install.sh | sh
2

Create Project

mkdir workshop-solutions cd workshop-solutions uv init
3

Add Dependencies

uv add crewai==0.114.0 uv add python-dotenv uv add streamlit
4

Activate Virtual Environment

source .venv/bin/activate
5

Set Environment Variables

Add your API keys as environment variables. Choose one of the following options:

Option 1: Create a .env file

echo "OPENAI_API_KEY=your_openai_api_key_here" > .env echo "SERPER_API_KEY=your_serper_api_key_here" >> .env

Option 2: Set directly in terminal (macOS/Linux)

export OPENAI_API_KEY='your_openai_api_key_here' export SERPER_API_KEY='your_serper_api_key_here'
Using Different LLM Providers?

If you're using an alternative AI model provider instead of OpenAI, please refer to the CrewAI LLM Connections documentation for specific setup instructions for your chosen provider.

Assignments

1

Your First Agent

Learn how to create and deploy your first AI agent using CrewAI, from basic setup to simple task automation.

View Details
2

Content Creation with Guardrails

Build an AI system that generates content based on user input while ensuring it adheres to safety guidelines, avoiding violence and inappropriate content.

View Details
3

Fraud Detection Workflow

Develop an AI system that analyzes and classifies transactions to identify potential fraudulent activities based on transaction patterns and behaviors.

View Details
4

AI-Powered Onboarding System

Create an intelligent onboarding system that personalizes the user experience using AI agents.

View Details
5

Final Project: AI-Powered Interactive System

🎯 Final Challenge: Build a comprehensive AI-powered interactive system using CrewAI and Streamlit. This project will demonstrate your mastery of multi-agent orchestration, real-time user interaction, and building production-ready applications with chat interfaces, voice input, and intelligent decision-making.

View Details