youssefHosni/LLM-Open-University-From-Begineer-to-Advanced

What it solves

Large Language Models (LLMs) are everywhere now — chatbots, search, coding assistants, agents, and more. But learning them can be overwhelming because the field is huge and scattered across blog posts, papers, videos, and courses. This repository organizes the best free resources into a single, structured path so that anyone can go from not knowing what a transformer is to building and deploying real LLM applications without getting lost.

How it works

The roadmap is divided into four sections that build on each other. The first section covers LLM basics: architecture, transformers, attention mechanisms, tokenization, embeddings, and decoding strategies. It points to classic explainers like Jay Alammar's illustrated Transformer and Andrej Karpathy's video series. The second section moves into building and training LLMs from scratch — dataset preparation, instruction tuning, fine-tuning (including parameter-efficient methods like LoRA and QLoRA), evaluation, quantization, RLHF alignment, and how to stay current with research. The third section shifts to production: prompt engineering, vector databases, RAG, LLM agents, inference optimization, LLMOps, security, deployment, local model running tools, and MCP. The final section is about portfolio projects — ideas and guided work so learners can show what they have built.

Who it's for

It's designed for absolute beginners who want to master LLMs step by step, as well as data scientists, machine learning engineers, AI engineers, and software developers who want to deepen their LLM skills. Because every section links only to free resources (articles, YouTube videos, courses), it's also for self-learners on a budget.

Highlights

• Structured, staged curriculum that moves logically from fundamentals to training to deployment to portfolio work
• Curated list of high-quality, mostly free resources: illustrated articles, Karpathy's and 3Blue1Brown's videos, university-style courses, and practical notebooks
• Includes 14 free Colab notebooks for hands-on fine-tuning practice
• Covers both technical depth (attention, PEFT, quantization, RLHF) and practical production topics (RAG, agents, LLMOps, security)
• Ends with portfolio-building projects so you can demonstrate your skills to employers