The LLM Development Stack

This is the official GitHub repository for the Pearson Live Training course “The LLM Development Stack”, created by Rob Barton and Jerome Henry.

This course introduces developers and engineers to the complete modern LLM development stack, covering foundational tools, local model-serving platforms, agentic frameworks and protocols, cloud-based deployment, and evaluation/benchmarking—all delivered through practical demos, code labs, and live Q&A.

---

Authors

Name	Role
Rob Barton	Cisco Distinguished AI Engineer
Jerome Henry	Cisco Distinguished Engineer

---

Course Outline

Section 1: LLM Development Foundational Tools

This section is a combination of demo and slides, introducing the environments and platforms used throughout the course.

• Google Colab
• Hugging Face
• Q&A

---

Section 2: LLM Management Platforms

Learn how to run LLMs locally using modern tools for inference and UI.

• Overview of Ollama and LMStudio
• Ollama deep dive
• Developing GenAI apps using Ollama and Open WebUI
• Q&A

---

Section 3: Agentic AI Development Tools

Build intelligent AI agents using Python and agentic development frameworks.

• Agentic frameworks: LangChain/LangGraph, CrewAI, AutoGen
• Automating agents with LangChain
• Programming your own agent with Python
• Q&A

---

Section 4: Agentic AI Development Protocols

Explore how agents interact and share context using cutting-edge protocols.

• Agentic Protocols
• Message Chain Protocol (MCP)
• Agent to Agent (A2A)
• Model Context Protocol (MCP)
• Q&A

---

Section 5: Cloud-based LLM Development Platforms

Deploy and scale your AI applications using cloud-based LLM tools.

• Overview of Cloud LLM development tools (Google Vertex, Azure AI Studio, AWS Bedrock)
• AWS Bedrock deep dive
• Building an AI Assistant to use Bedrock
• Q&A

---

Section 6: LLM Monitoring and Observability

Learn how to monitor and observe LLM-based applications.

• Overview of LLM monitoring and observability
• An introduction to LangSmith
• Q&A

---

Prerequisites

To make the most of this course, you should have:

• Basic Python programming experience
• Familiarity with AI/ML concepts
• A GitHub account and the ability to use browser-based coding tools (e.g., Google Colab)

---

Repository Structure

The repository is organized into sections that correspond to the course modules, with additional supporting files:

Course Sections

• Lesson-1-Foundational-Tools/ – Includes Google Colab and Hugging Face demos to introduce the development environment
• Lesson-2-LLM-Management-Platforms/ – Resources and examples using Ollama, LM Studio, and Open WebUI
• Lesson-3-Agentic-Development-Tools/ – Agentic AI frameworks like LangChain, LangGraph, CrewAI, and AutoGen
• Lesson-4-Agentic-Protocols/ – Covers Message Chain Protocol (MCP), Agent-to-Agent (A2A), and Model Context Protocol
• Lesson-5-Cloud-LLM-Platforms/ – Hands-on projects using AWS Bedrock, Google Vertex, and Azure AI Studio
• Lesson-6-LLM-Monitoring and Observability/ – An introduction to LangSmith

Supporting Files

• requirements.txt – Lists all Python package dependencies
• .gitignore – Specifies files and directories to be ignored by Git

Each section directory typically contains:

• Python scripts (.py) and Jupyter notebooks (.ipynb) with working examples
• Data files used in the examples (where applicable)
• README files with section-specific instructions
• Visualizations and demos
• Additional explanatory materials

Note: Some sections may be added or updated as the course progresses.

---

Getting Started

1. Clone this repository to your local machine:

   git clone https://github.com/robbarto2/Oreilly-LLM-Development-Stack.git

2. Navigate to the project directory:

   cd Oreilly-LLM-Development-Stack

3. Install dependencies:

   pip install -r requirements.txt

4. Open any of the demo notebooks in Google Colab or your preferred IDE.

5. (Optional) For local development, install tools like Ollama and LM Studio.

---

This repository will be updated regularly as the course progresses.