AI Agents with LangGraph Course - My Learning Journey

In this course, I learned to build advanced AI agents using LangGraph, an extension of the popular LangChain framework.

What I Learned

Through hands-on practice, I gained skills in:

• Building AI agents from scratch using Python and LLMs
• Implementing agents with LangGraph's component architecture
• Enhancing agent capabilities through agentic search
• Managing state and persistence in agent systems
• Incorporating human feedback loops
• Creating specialized agents for complex tasks

Course Content & Key Takeaways

1. Building Agents from Scratch

• Built a basic agent architecture from the ground up
• Mastered the interaction between LLMs and supporting code
• Implemented in: 01_simple_react_agent_from_scratch.ipynb

2. LangGraph Components

• Explored core components and architectural patterns
• Created agents using LangGraph's framework
• Developed in: 02_langraph_components.ipynb

3. Agentic Search

• Integrated advanced search capabilities
• Enhanced agent knowledge through multiple data sources
• Demonstrated in: 03_agentic_search.ipynb

4. Persistence and Streaming

• Implemented state management across conversation threads
• Developed robust conversation handling systems
• Applied in: 04_persistence_and_streaming.ipynb

5. Human-in-the-Loop

• Built interactive agent systems
• Integrated user feedback mechanisms
• Showcased in: 05_human_in_the_loop.ipynb

6. Essay Writing Agent

• Created a practical research and writing agent
• Replicated human researcher workflows
• Completed in: 06_essay_writer.ipynb

Technical Setup

Prerequisites Used

• Python
• Understanding of LLMs and AI concepts
• LangChain & LangGraph framework

Environment Setup

1. Installed required packages:

pip install openai langchain langgraph tavily-python langchain-openai langchain-community

2. Configured environment:

• Created .env file from .env.template
• Set up API keys and configurations

Project Structure

Each notebook contains complete implementations with explanations and working code examples. The course follows a logical progression from basic concepts to advanced applications.

Instructors

Learned directly from industry experts:

• Harrison Chase (LangChain founder)
• Rotem Weiss (Tavily founder)