The transition from simple LLM chatting to building autonomous AI agents is the biggest shift in software engineering today. While the field moves fast, these 10 repositories represent the “Gold Standard” of learning resources, offering everything from foundational math to production-grade deployment patterns.
Foundations and Frameworks
1. Hands-On Large Language Models
The Deep Dive: This is the definitive companion for understanding the “brain” behind the agent. It provides complete code notebooks covering the full lifecycle of an LLM.
- Key Focus: Fine-tuning, quantization, and evaluation.
- Direct Link: github.com/Hands-On-LLMs/Hands-On-LLMs
2. AI Agents for Beginners (Microsoft)
The Starting Line: Microsoft’s 11-part course is the best structured entry point. It focuses on the logic of agency—how a model uses tools to solve multi-step problems.
- Key Focus: Reasoning loops and tool-calling basics.
- Direct Link: github.com/microsoft/ai-agents-for-beginners
3. GenAI Agents
The Practical Sandbox: This repo bridges the gap between creative generation and functional utility. It’s ideal for those building agents that need to interact with diverse media and APIs.
- Key Focus: Practical generative workflows.
- Direct Link: github.com/instill-ai/genai-agents
Engineering and Deployment
4. Made with ML
Production Focus: This is widely considered one of the best resources for learning MLOps. It teaches you how to design and deploy apps that are reliable enough for the real world.
- Key Focus: Testing, versioning, and deployment pipelines.
- Direct Link: github.com/GokuMohandas/Made-With-ML
5. Prompt Engineering Guide
The Language of Agents: Agents are only as good as their instructions. This guide covers advanced techniques like Chain-of-Thought and ReAct—the bread and butter of agentic reasoning.
- Key Focus: Optimization and prompt reliability.
- Direct Link: github.com/dair-ai/Prompt-Engineering-Guide
6. Hands-On AI Engineering
Applied Use Cases: If you learn best by doing, this repo provides end-to-end examples of LLM-powered applications that go beyond simple “Hello World” scripts.
- Key Focus: Real-world agent architecture.
- Direct Link: github.com/Sumanth077/Hands-On-AI-Engineering
Theory and System Design
7. Awesome Generative AI Guide
The Knowledge Hub: A curated megathread of research and tools. It’s the “Home Base” for anyone who needs to stay on the cutting edge of SOTA (State of the Art) research.
- Direct Link: github.com/steven2358/awesome-generative-ai
8. Designing Machine Learning Systems
The Architect’s Blueprint: Agents are complex systems. This repo (based on Chip Huyen’s work) teaches you how to think about data flow, monitoring, and system architecture.
- Direct Link: github.com/chiphuyen/machine-learning-systems-design
9. ML for Beginners (Microsoft)
The Fundamentals: You can’t build advanced agents without a solid grasp of classic ML. This 12-week curriculum is the best way to shore up your foundational knowledge.
- Direct Link: github.com/microsoft/ML-For-Beginners
10. LLM Course
The Roadmap: This repo provides a clear, tiered path for three distinct roles: the Scientist, the Engineer, and the Beginner. It is arguably the most comprehensive roadmap on GitHub.
- Direct Link: github.com/mlabonne/llm-course
