🧠 Learn LLMs — A Complete Practical Course

From zero to building real AI applications with Large Language Models.

---

📚 What Is This?

This is a free, open-source course on Large Language Models (LLMs) — designed for everyone from complete beginners to developers who want to build production-grade AI apps.

No fluff. Just clean explanations, working code, and real examples.

---

🗺️ Course Roadmap

📦 llm-course
 ┣ 📂 01-introduction              → What are LLMs? How do they work?
 ┣ 📂 02-transformer-architecture  → How transformers work under the hood
 ┣ 📂 03-prompt-engineering        → Craft prompts like a pro
 ┣ 📂 04-llm-apis                  → Use OpenAI, Claude, Gemini APIs
 ┣ 📂 05-rag                       → Build retrieval-augmented generation apps
 ┗ 📂 06-fine-tuning               → Fine-tune models on your own data

---

🧭 Who Is This For?

👤 Profile	✅ You'll Learn
Beginners	What LLMs are, how to prompt them, use APIs
Developers	Build RAG pipelines, integrate LLM APIs into apps
ML Engineers	Fine-tune models, evaluate outputs, go to production

---

📖 Modules

🟢 Module 1 — Introduction to LLMs

Perfect for beginners. No code required.

• What is an LLM?
• How does training work? (Tokens, transformers, attention)
• Famous models: GPT-4, Claude, Gemini, Mistral, LLaMA
• Key concepts: Temperature, context window, hallucinations

📁 01-introduction/

---

🟡 Module 2 — Transformer Architecture

Understand what's actually happening inside LLMs.

• Tokenization & embeddings
• Queries, Keys, Values — how attention works
• Multi-head attention with math + diagrams
• Feed-forward networks, residuals, layer norm
• How it all stacks into a full model
• Build a mini GPT from scratch (~150 lines of PyTorch)

📁 02-transformer-architecture/

---

🟠 Module 3 — Prompt Engineering

The skill that 10x's your LLM output quality.

• Zero-shot, few-shot, chain-of-thought prompting
• Role prompting & system prompts
• Prompt templates & reusability
• Common mistakes and how to fix them

📁 02-prompt-engineering/

---

🟠 Module 3 — LLM APIs

Connect your apps to powerful AI models.

• OpenAI API (GPT-4o, GPT-3.5)
• Anthropic Claude API
• Google Gemini API
• Open-source models via Ollama & HuggingFace
• Building a simple chatbot

📁 03-llm-apis/

---

🔵 Module 4 — RAG (Retrieval Augmented Generation)

Give your LLM access to your own data.

• Why RAG? The problem with LLM memory
• Embeddings & vector databases (FAISS, Pinecone, ChromaDB)
• Build a RAG pipeline from scratch
• Chat with your PDF

📁 04-rag/

---

🔴 Module 5 — Fine-Tuning LLMs

Train a model on your own data.

• When to fine-tune vs. prompt engineer vs. RAG
• Preparing your dataset (JSONL format)
• Fine-tuning with OpenAI & HuggingFace
• Evaluating your fine-tuned model

📁 05-fine-tuning/

---

🚀 Quick Start

# Clone the repo
git clone https://github.com/YOUR_USERNAME/llm-course.git
cd llm-course

# Install Python dependencies
pip install -r requirements.txt

# Set your API keys
cp .env.example .env
# → Edit .env and add your keys

---

🔧 Prerequisites

• Python 3.9+
• Basic Python knowledge (for code modules)
• API keys: OpenAI | Anthropic (free tiers available)

---

🤝 Contributing

Contributions are welcome! Found a bug, have a better example, or want to add a new module?

1. Fork the repo
2. Create a branch: git checkout -b feature/your-idea
3. Commit and push
4. Open a Pull Request

See CONTRIBUTING.md for details.

---

📄 License

MIT License — free to use, share, and build on.

---

⭐ Star This Repo

If this helped you, give it a star! It helps others find it.

---

Made with ❤️ to make AI education accessible to everyone.