Reference · Syllabus

Syllabus

30 lessons in 8 parts, mapped node by node to roadmap.sh/inference-engineering. Built for an app developer becoming an inference engineer in 13 weeks at about 5 hours a week.

The mission

Become hireable as an inference engineer. By the end you can do the napkin math (TTFT, ITL, KV cache, $/1M tokens), serve and load-test models with vLLM and llama.cpp, apply and measure the core optimizations, and show a public capstone repo. The full mission is in MISSION.md.

How every lesson works

  1. Concept, briefly. Only the knowledge the lab needs, with sources in the margin.
  2. Predict. You write down what you expect before running anything, because a wrong prediction teaches more than a right one.
  3. Lab. Hands-on on your Mac or a free cloud GPU, with step-by-step checklists and a results checker.
  4. Exercises. Quizzes and napkin-math problems with instant feedback, plus one "explain it" answer you write from memory.
  5. Recall. From Lesson 2 on, a few questions from earlier lessons, because spacing is what makes knowledge stick.
  6. Portfolio. Every lab adds a result to portfolio/LOG.md, the raw material for your capstone and interviews.

Lessons are generated one at a time. Finish one, tell your teacher how it went, and the next one is written for where you actually are.

Why the order differs from the roadmap

Lab platforms: Mac your M1 Pro · Colab T4 free NVIDIA GPU in the browser · Kaggle 2×T4 free, two GPUs · Calculator / sim runs anywhere.

Roadmap coverage

Every node on the roadmap and the lesson that teaches it: .