If you find this repository helpful, please consider giving it a ⭐ to help other learners discover it!
A comprehensive, from-scratch guide to mastering Multi-Agent Systems (MAS) and Agentic Architectures.
Welcome to the Agentic Design Patterns repository!
I built this project during my own learning journey to help fellow developers and AI enthusiasts understand complex Multi-Agent Systems and Agentic RAG architectures from the ground up.
Whether you are just starting with basic ReAct agents or building complex, hierarchical, debate-driven multi-agent ecosystems, this repository breaks down each concept into digestible, interactive tutorials. Every pattern includes:
- 📖 Comprehensive Markdown Documentation (Theory, Pros/Cons, Use Cases)
- 🐍 Clean Python Implementations (Built from scratch using LangGraph)
- 📓 Interactive Jupyter Notebooks (For hands-on execution and experimentation)
Click on the folder emoji (📂) to explore the code, documentation, and notebooks for each architectural pattern.
| S.No | Architecture Name | Explore |
|---|---|---|
| 1 | ReAct (Reason + Act) | 📂 |
| 2 | Self-Reflection / Reflexion | 📂 |
| 3 | Plan and Execute | 📂 |
| 4 | Human-In-The-Loop | 📂 |
| 5 | Multi-Agent Systems Overview | 📂 |
| 6 | Network (Swarm) Multi-Agent System | 📂 |
| 7 | Hierarchical Multi-Agent System | 📂 |
| 8 | Parallel Map-Reduce Multi-Agent System | 📂 |
| 9 | Supervisor-As-Tools Multi-Agent System | 📂 |
| 10 | Agentic RAG | 📂 |
| 11 | Corrective RAG (Corrective-Rag-Paradigm) | 📂 |
| 12 | Self RAG (Self-Rag-Paradigm) | 📂 |
| 13 | Multi-Agentic RAG Overview | 📂 |
| 14 | Hierarchical RAG (Multi-Agent Paradigm) | 📂 |
| 15 | Sequential Multi-Hop RAG (Multi-Agent Paradigm) | 📂 |
| 16 | Collaborative Debate RAG (Multi-Agent Paradigm) | 📂 |
| 17 | Adaptive RAG (Adaptive-Rag-Paradigm) | 📂 |
| 18 | Autonomous RAG (Autonomous-Rag-Paradigm) | 📂 |
- PDF Pattern Notes: I will be compiling a comprehensive, detailed PDF study guide covering all these Agentic Patterns in the near future. Stay tuned!
- Open-Source Contributions: This repository is open to the community! If you are learning or building other agentic paradigms (e.g., Swarms, routing tricks, multi-modal agents) and want to add them as a pattern, you are highly welcome to open a Pull Request. Let's make this the ultimate learning hub for Agentic AI!
- Clone the repo:
git clone https://github.com/AsimShah/Agentic-Design-Patterns.git cd Agentic-Design-Patterns - Set up the Virtual Environment:
All dependencies are listed in
requirements.txt.python -m venv virtEnvAgenticPatterns source virtEnvAgenticPatterns/bin/activate # On Windows: virtEnvAgenticPatterns\Scripts\activate pip install -r requirements.txt
- Configure Environment Variables:
Create a
.envfile in the root directory and add your API keys (e.g.,OPENAI_API_KEY,GROQ_API_KEY). - Explore: Open the Jupyter Notebooks inside each pattern's folder to run the agents step-by-step!
- Enhance Markdown Viewing:
To view the workflow flowcharts and state machines, it is highly recommended to view the markdown files (
.md) using an extension that supports Mermaid.js rendering.- For VS Code: Install the Markdown Preview Enhanced extension or the Mermaid Preview extension.
- For GitHub: Native rendering is supported out of the box!
Built with ❤️ by Asim Shah
I am passionate about Artificial Intelligence, LLMs, and building robust, autonomous agentic systems. I created this repository to bridge the gap between theoretical AI concepts and practical, production-ready code.
Let's connect and build the future of AI together!
- 💼 LinkedIn: Connect with me
- 🐙 GitHub: Follow my projects
If you find this repository helpful, please consider giving it a ⭐ to help other learners discover it!