LangGraph AI Agents Mastery 🚀

This repository is a comprehensive guide and portfolio project based on the "AI Agents in LangGraph" course by DeepLearning.AI, featuring Harrison Chase (LangChain) and Rotem Weiss (Tavily). It explores the transition from simple LLM chains to complex, stateful, and cyclic agentic workflows.

🛠 Project Structure

The project is organized into progressive modules:

• 01_Basics/: Building a ReAct agent from scratch and introducing LangGraph components (Nodes, Edges, State).
• 02_State_Management/: Advanced persistence using SQLite, conversation threads, and real-time token streaming.
• 03_Tool_Integration/: Leveraging Agentic Search (Tavily) for LLM-optimized information retrieval.
• 04_Human_in_the_loop/: Implementing human-approval gates, state editing, and "Time Travel" debugging.
• 05_Use_Cases/: A full-scale Essay Writer Agent utilizing a cyclic Reflection workflow.

🌟 Key Concepts Implemented

1. Agentic Workflows

Unlike linear chains, these workflows are iterative. Agents plan, act, reflect, and use tools to track progress over multiple steps.

2. State Management & Persistence

Using Annotated types and SqliteSaver to give agents long-term memory and the ability to resume conversations across different threads.

3. Human-in-the-Loop

Strategic interrupt_before points that allow humans to approve or modify agent actions (e.g., confirming a search query or a financial transaction).

4. Time Travel (State Manipulation)

The ability to go back in history, fork the state, and re-run agentic logic from a specific point in time for debugging and steering.

🚀 Getting Started

1. Clone the repo:

   git clone [https://github.com/ozereray/langgraph-agentic-workflows]
   

2. Install dependencies:

   pip install -r requirements.txt

3. Set up environment variables: Create a .env file and add:

   OPENAI_API_KEY=your_key
   TAVILY_API_KEY=your_key

📚 Acknowledgments

Special thanks to Harrison Chase and Rotem Weiss for the amazing insights provided in the DeepLearning.AI course.