03_The_Agent_Loop

Session 3: The Agent Loop

Quicklinks

📰 Session Sheet	⏺️ Recording	🖼️ Slides	👨‍💻 Repo	📝 Homework	📁 Feedback
The Agent Loop	Recording! passcode: zD6.J*fM	Session 3 Slides	You are here!	Session 3 Assignment: Agentic Loop	Feedback 1/20

Outline:

🤜 BREAKOUT ROOM #1:

• Task 1: Dependencies
• Task 2: Environment Variables
• Task 3: LangChain Core Concepts (Runnables & LCEL)
• Task 4: Understanding the Agent Loop
• Task 5: Building Your First Agent with create_agent()
  • 🏗️ ACTIVITY #1: Create a Custom Tool

🤜 BREAKOUT ROOM #2:

• Task 6: Loading & Chunking Documents
• Task 7: Setting up Qdrant Vector Database
• Task 8: Creating a RAG Tool
• Task 9: Introduction to Middleware
• Task 10: Building Agentic RAG with Middleware
  • 🏗️ ACTIVITY #2: Enhance the Agent

Steps to Run:

1. Install UV, which you can do through this resource
2. Run the command uv sync
3. Open your Jupyter notebook and select .venv for your kernel.

Build 🏗️

Run the notebook!

Ship 🚢

• Add one of the following enhancements (or whatever augmentations suit your use case) to the agentic RAG system:
  • Add a new tool (BMI calculator, calorie estimator, etc.)
  • Create custom middleware (logging, guardrails, rate limiting)
  • Improve the RAG tool (metadata filtering, reranking, citations)
  • Add conversation memory to the agent
• Make a simple diagram of your agent architecture
• Run the notebook
• When you're finished with augmentations, compare your enhanced agent to the baseline!
• Record a Loom video walking through the notebook, the questions in the notebook, and your addition!

Share 🚀

• Show your Agent in a Loom video and explain the diagram
• Make a social media post about your final application and tag @AIMakerspace
• Share 3 lessons learned
• Share 3 lessons not learned

Here's a template to get your post started!

🚀 Exciting News! 🎉

I just built my first AI Agent using LangChain 1.0's create_agent API! 🤖💼

🔍 Three Key Takeaways:
1️⃣ The agent loop is the foundation of all AI agents - understanding it unlocks so much potential! 🧠✨
2️⃣ Middleware lets you hook into every step of the agent loop for logging, guardrails, and more! 🔧📈
3️⃣ Agentic RAG gives agents control over when to retrieve - way more flexible than traditional RAG! 🔄📚

A huge shoutout to @AI Makerspace for their invaluable resources and guidance. 🙌

Looking forward to more agentic adventures! 🌟 Feel free to connect if you'd like to chat more about it! 🤝

#AI #Agents #LangChain #BuildInPublic

🚧 Advanced Build (Optional): Human-in-the-Loop Middleware

Implement a Wellness RAG Agent with Human-in-the-Loop Middleware.

• Use HumanInTheLoopMiddleware to pause agent execution before tool calls
• Create a custom approval function that lets users approve, reject, or edit tool inputs
• Add the middleware to your wellness agent
• Test with queries that trigger different tools

This pattern is critical for production systems requiring human oversight!

NOTE: Can be completed in lieu of full notebook

Submitting Your Homework

Main Assignment

Follow these steps to prepare and submit your homework:

1. Pull the latest updates from upstream into the main branch of your AIE9 repo:
   • (You should have completed this process already.) For your initial repo setup, see Initial_Setup
   • To get the latest updates from AI Makerspace into your own AIE9 repo, run the following commands:

   git checkout main
   git pull upstream main
   git push origin main
   

2. IMPORTANT: Start Cursor from the 03_The_Agent_Loop folder (you can also use the File -> Open Folder menu option of an existing Cursor window)
3. Answer Questions 1 - 4 using the ##### ✅ Answer: markdown cell below them.
4. Complete Activity #1 and Activity #2 in the notebook.
5. Add, commit and push your modified The_Agent_Loop_Assignment.ipynb to your GitHub repository.

When submitting your homework, provide:

• Your Loom video link
• The GitHub URL to your completed notebook