Agentic AI Projects

This repository contains projects developed as part of The Complete Agentic AI Engineering Course on Udemy.

Projects

01_deep_research

A multi-agent system that performs deep research on a given query:

• Uses a research manager to coordinate the process
• Includes specialized agents for search, planning, writing, and email communication
• Generates comprehensive research reports and can send them via email
• Framework: LangChain + OpenAI

02_debate

A debate system built with crewAI:

• Creates a structured debate between multiple AI agents on specified topics
• Agents have different perspectives and roles in the debate
• Produces a comprehensive report with arguments and conclusions
• Framework: CrewAI

03_financial_researcher

A financial research assistant built with crewAI:

• Analyzes financial data and markets
• Provides insights on investments and financial trends
• Generates detailed financial reports
• Framework: CrewAI

04_stock_picker

A stock analysis and recommendation system:

• Researches and analyzes stocks
• Maintains memory of past analyses
• Generates investment recommendations and reports
• Framework: CrewAI

05_coder

An AI coding assistant:

• Helps with software development tasks
• Can generate code based on requirements
• Assists with debugging and code improvements
• Framework: CrewAI

06_engineering_team

A simulated engineering team of AI agents:

• Collaborates on software development projects
• Different agents take on different engineering roles
• Handles the full development lifecycle
• Framework: CrewAI

07_sidekick

An AI assistant with multiple capabilities that can help you complete tasks using various tools:

• Web browsing - can navigate websites and extract information
• File management - can create, read, and modify files
• Python execution - can write and run Python code
• Push notifications - can send notifications to keep you updated
• Wikipedia access - can search and retrieve information from Wikipedia
• Web search - can perform online searches for information
• Framework: LangChain + LangGraph

Setup

Each project has its own dependency requirements. Generally, you'll need to:

1. Create a Python environment
2. Install the required dependencies
3. Set up the necessary API keys in .env files

For crewAI-based projects (02_debate, 03_financial_researcher, etc.), use:

pip install uv
crewai install

For other projects, install dependencies using pip or a package manager specified in the project directory.

Running the Projects

Each project has its own run instructions. For example:

• Sidekick: Navigate to 07_sidekick and run python app.py
• CrewAI projects: Run crewai run from the project directory

Dependencies

The main dependencies for the projects include:

• LangChain & LangGraph: For creating AI agent workflows (01_deep_research, 07_sidekick)
• CrewAI: For multi-agent systems (02_debate, 03_financial_researcher, 04_stock_picker, 05_coder, 06_engineering_team)
• Gradio: For building user interfaces
• Playwright: For web browsing capabilities
• Various LLM APIs (OpenAI, Anthropic, etc.)

See the pyproject.toml file in each project directory for specific dependencies.

Course Information

These projects are based on The Complete Agentic AI Engineering Course on Udemy, which covers:

• Building autonomous AI agents
• Multi-agent systems
• Tool use and web browsing
• Agent memory and reasoning
• LangChain, LangGraph, and CrewAI frameworks