Welcome to the Agentic AI Tutorial, a comprehensive guide designed to empower developers in creating sophisticated agentic AI systems. This resource delves into autonomous agents capable of reasoning, planning, and executing actions through the use of advanced Large Language Models (LLMs). Spanning from basic LLM interactions to the construction of fully autonomous agents, this tutorial caters to both beginners and intermediate developers.

Why Choose Agentic AI?

Agentic AI transcends traditional AI models that merely respond to prompts by introducing key functionalities such as:

• Autonomy: The ability to select appropriate tools and methods for problem-solving.
• Reasoning: The capability to decompose complex tasks into achievable steps.
• Persistence: Maintenance of state and memory throughout lengthy interactions.
• Action: The capacity to engage with the real world using APIs, databases, and files.

Learning Journey

The tutorial is structured into a roadmap comprising several chapters:

Core Technology Stack

The tutorial employs a robust technology stack that includes:

• Frameworks: LangChain, LangGraph
• Models: OpenAI (GPT-4o), Google Gemini (2.0 Flash), Ollama (Local Llama 3/Mistral)
• Vector Databases: Chroma, FAISS
• Embeddings: Sentence Transformers (HuggingFace)

Getting Started

To begin the tutorial, a few prerequisites are necessary:

• An installation of Python 3.8 or greater.
• API Keys for OpenAI/Google (optional if utilizing only Ollama).

Each chapter of the tutorial is equipped with examples and detailed explanations that foster a deeper understanding of the concepts presented.

Chapters Overview

• Chapter 1: LLM Fundamentals: Explore direct API interactions and system prompt engineering.
• Chapter 2: LangChain Orchestration: Learn about LCEL, as well as building various chains and routing methods.
• Chapter 3: Memory & Context: Discuss memory management techniques and retrieval-augmented generation with local vector stores.
• Chapter 4: Autonomous Agents: Understand state graphs, collaboration among multiple agents, and safe production practices.

Contribution Guidelines

Contributions are welcomed, whether it’s fixing minor issues or introducing new features. Steps for contributing include:

1. Fork the project.
2. Create a feature branch.
3. Commit changes with a descriptive message.
4. Push the branch to GitHub.
5. Open a pull request for review.

Author Information

Zkzk - AI Engineer & Educator. Contributions and inquiries can reach through GitHub.

_Disclaimer: This tutorial is intended for educational use. Users should be aware that costs may arise from cloud LLM usage.