# mlabonne/llm-course — LLM Engineer Track

**Type:** article / GitHub curriculum (educational)
**Tier:** 4 (Educational)
**Author(s):** Maxime Labonne
**Date:** Accessed 2026-06-01
**URL:** https://github.com/mlabonne/llm-course
**Accessed:** 2026-06-01

## Why This Source Matters

A highly-starred, frequently updated open curriculum that organizes the practitioner's path into three tracks. Its "LLM Engineer" track maps almost one-to-one onto this lab's project sequence, which validates the curriculum ordering and provides a cross-reference for "running LLMs" as the foundational first step (Project 1).

## Key Claims

- The course is organized into three parts: **LLM Fundamentals** (math, Python, neural nets), **The LLM Scientist** (building optimal LLMs), and **The LLM Engineer** (creating and deploying applications).
- The **LLM Engineer** track's first topic is **"Running LLMs"**: accessing models via APIs (OpenAI, Anthropic, Hugging Face) or locally (LM Studio, Ollama); prompt engineering (zero-shot, few-shot, chain-of-thought); structured output.
- Subsequent topics: vector storage & RAG → retrieval-augmented generation → advanced RAG & agents → inference optimization (Flash Attention, KV caching, speculative decoding) → deployment → security (prompt injection, data poisoning).
- The track emphasizes application development over model training.

## Relevant To

- concepts: [running-llms, api-access, prompt-engineering, curriculum-sequencing]
- projects: [01-ai-chatbot, 03-semantic-search, 04-pdf-research-assistant, 08-ai-agent]

## Notes

- Use as a curriculum cross-reference and for breadth, not for authoritative API details. "Running LLMs" being step 1 of the Engineer track corroborates Project 1's position as the foundation.
- Related curated lists gathered the same session: github.com/HandsOnLLM/Hands-On-Large-Language-Models, github.com/dair-ai/Prompt-Engineering-Guide, github.com/rohitg00/ai-engineering-from-scratch (Phase 12 = LLM Engineering maps to Projects 1–4).