🎓 LLM Learning Path — From Zero to Researcher

A structured, self-paced curriculum for mastering Large Language Models and LLM-based Agents.

Timeline: ~5 months (can be shortened to ~3 months with prior DL experience)

📋 Curriculum Overview

Phase	Topic	Duration	Status
1	Foundations	2-3 weeks	⬜
2	Transformers	2-3 weeks	⬜
3	Pre-training & Scaling	2-3 weeks	⬜
4	Fine-Tuning & Alignment	2-3 weeks	⬜
5	Inference & Deployment	1-2 weeks	⬜
6	Prompting & Reasoning	1-2 weeks	⬜
7	LLM Agents	2-4 weeks	⬜
8	Advanced Research	Ongoing	⬜

🗺️ How to Use This Repo

1. Go phase by phase — each folder has its own README with objectives, readings, and exercises
2. Check off items as you complete them (edit the checkboxes in each phase)
3. Take notes in the notes/ folder — one file per phase
4. Do the exercises — hands-on work is where the real learning happens
5. Track papers you've read in the reading log

📐 Prerequisites

• Python programming (intermediate+)
• Basic linear algebra (vectors, matrices, dot products)
• Basic calculus (derivatives, chain rule)
• Basic probability (distributions, Bayes' theorem)

🔑 Key Resources (Quick Access)

Resource	Type	Link
Karpathy — NN: Zero to Hero	Video Series	YouTube
HuggingFace LLM Course	Course	HF Learn
Lilian Weng Blog	Blog	lilianweng.github.io
Papers With Code	Reference	paperswithcode.com
arXiv cs.CL	Papers	arxiv.org/list/cs.CL

📜 License

This learning path is open source. Feel free to fork, modify, and share.