Mastering AI Agents: Advanced Patterns & Architectures

Welcome to the Mastering AI Agents workshop. This is a practical, code-first curriculum designed to help you build sophisticated, reliable autonomous systems from scratch using raw patterns—no high-level frameworks required.

🚀 Overview

Modern AI development is moving from simple prompts to complex agentic workflows. This workshop covers 45+ patterns identified in research and real-world implementations, grouped into 7 practical chapters.

📚 Curriculum

Chapter	Focus	Key Patterns
1. Building Blocks	Foundations of Agency	Chaining, Routing, Parallelization
2. Reasoning Loops	Solving Complex Logic	CoT, Self-Correction, Decomposition
3. Taking Action	Tool Use & Execution	ReAct, Planning, Code Execution
4. Self-Correction	Reliability & Safety	Reflection, Guardrails, Pydantic
5. Memory & Context	Long-term Intelligence	Episodic/Semantic Memory, Graph Memory
6. Collaboration	Multi-Agent Systems	Multi-Agent, Blackboard, Human-in-the-loop
7. Mastering Agency	Advanced Orchestration	Agentic RAG, Recursive Agents, Meta-Controller

🛠️ Setup

1. Prerequisites

• Python 3.12+
• uv (recommended) or pip
• Docker (optional)

2. Environment Setup

cp env.example .env
# Open .env and add your API keys (Google Gemini, OpenAI, etc.)

3. Install Dependencies

make setup

4. Start Workshop

make dev

Navigate to http://localhost:8888 in your browser.

🧪 Philosophy

We believe in learning the raw patterns before reaching for a framework. By understanding the underlying logic of agentic workflows, you can build more flexible, debuggable, and performant systems.

---

Built with ❤️ by LearnWithParam