How to Find Your Starting Level in the AI Engineering from Scratch Curriculum

TLDR: The rohitg00/ai-engineering-from-scratch repository includes a built-in placement quiz called /find-your-level that scores your AI/ML knowledge across ten questions and maps your result to the optimal curriculum phase.

If you are unsure where to begin, the AI Engineering from Scratch curriculum provides an automated way to find your starting level. The project is structured across progressive phases, from foundational math to advanced agent engineering, and entering at the wrong point wastes review time. According to the source code, the /find-your-level agent skill evaluates your existing expertise and returns a personalized learning path using hour estimates from ROADMAP.md.

How the Placement Quiz Evaluates Your Knowledge

The /find-your-level skill administers ten multiple-choice questions that cover five core knowledge areas: Math & Statistics, Classical ML, Deep Learning, NLP & Transformers, and Applied AI. Each correct answer earns one point, producing a total score out of ten.

Your score directly determines your recommended entry phase:

• 0–3 points maps to Phase 1: Math Foundations
• 8–9 points maps to Phase 11: LLM Engineering
• 10 points maps to Phase 14: Agent Engineering

After tallying your score, the skill generates a markdown table that labels each phase as Skip, Review, or Do, paired with the estimated hours defined in ROADMAP.md.

Where the Placement Skill Is Defined

The quiz logic is defined as an agent skill rather than a standalone script. The primary definition lives in .claude/skills/find-your-level/SKILL.md, which stores the full question flow, scoring rules, and personalized table generation logic.

The README.md introduces the curriculum and highlights the /find-your-level command so new users know it exists. Additionally, the web interface in site/lesson.html renders a call-out that instructs users to run /find-your-level inside supported agents.

How to Run the /find-your-level Quiz

You invoke the skill inside any compatible LLM agent, such as Claude, Cursor, Codex, OpenClaw, or Hermes. The skill is a markdown-defined flow that the agent interprets directly, so no local compilation is required.

If you have cloned the repository locally, you can point a Claude-compatible interpreter at the skill file:

# Clone the repo (optional if you just want to run the skill in an agent)

git clone https://github.com/rohitg00/ai-engineering-from-scratch.git
cd ai-engineering-from-scratch

# In a supported LLM agent, run the placement quiz

/find-your-level

During a typical interactive session, the agent presents questions one at a time, tallies your answers, and outputs the recommendation:

Agent: Welcome! Let’s find your level. Here’s the first question:

Q1. You have two vectors a=[1,2,3] and b=[4,5,6]. What is their dot product?
A) 21  B) 32  C) 15  D) 27
Your answer: B

... (questions 2-10) ...

Your total score: 7/10
Recommended entry point: Phase 7 – Transformers Deep Dive

Key Files Behind the Placement Logic

Several repository files cooperate to surface and execute the placement quiz:

• .claude/skills/find-your-level/SKILL.md — Contains the complete skill definition, including the ten-question quiz, scoring logic, and personalized learning-path generation.
• README.md — Introduces the curriculum and documents the /find-your-level command.
• site/lesson.html — Displays the “Run /find-your-level” call-out in the web UI for agent users.
• ROADMAP.md — Supplies the hour estimates used when the skill builds your personalized phase table.

Summary

• The AI Engineering from Scratch curriculum provides a /find-your-level skill that functions as a built-in placement quiz.
• Ten multiple-choice questions across five knowledge areas produce a score out of ten that maps to a recommended starting phase.
• The skill definition resides in .claude/skills/find-your-level/SKILL.md and is executed by compatible LLM agents.
• Results are rendered as a markdown table marking each phase as Skip, Review, or Do, with time estimates pulled from ROADMAP.md.

Frequently Asked Questions

What is the /find-your-level command in AI Engineering from Scratch?

/find-your-level is an agent skill defined in .claude/skills/find-your-level/SKILL.md that runs a ten-question placement quiz. It automatically scores your answers across five AI/ML domains and recommends the curriculum phase best suited to your current knowledge.

Which repository file contains the placement quiz logic?

The complete quiz flow, scoring rules, and personalized table generation are implemented in .claude/skills/find-your-level/SKILL.md. The README.md and site/lesson.html files only surface the command to users.

Can I run the placement quiz without a Claude-compatible agent?

The skill is engineered for LLM agents such as Claude, Cursor, Codex, OpenClaw, and Hermes. You can run it locally by feeding the SKILL.md file to a Claude-compatible interpreter, but the repository does not ship a standalone Python script.

How does the quiz determine my starting phase?

You earn one point per correct answer for a maximum score of ten. The skill translates point ranges into specific phases, for instance routing scores of 0–3 to Phase 1, 8–9 to Phase 11, and a perfect ten to Phase 14. It then references ROADMAP.md to display a clear table indicating which phases to skip, review, or complete.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}