How to Get Certified Through the Claude Certification Academy: A Complete Guide

The Claude Certification Academy provides free, self-paced certification tracks within the ai-engineering-from-scratch repository that require completing structured lessons, passing quizzes, executing labs via scripts/lesson_run.py, and verifying completion through scripts/audit_certifications.py.

The Claude Certification Academy is a formal credentialing system embedded in the rohitg00/ai-engineering-from-scratch open-source curriculum. It transforms the repository's standard lesson phases into certification tracks (CCDV-F, CCAR-P, CCAR-F, CCAO-F) that add diagnostic assessments, capstone projects, and automated verification steps required to earn a credential.

Understanding the Certification Structure

The academy operates as a structured subdirectory at certifications/claude/ inside the repository. Unlike the informal learning paths in the phases/ directory, certification tracks enforce a specific sequence of validated learning artifacts.

Track definitions live in certifications/claude/tracks/ as JSON files (e.g., ccar-f.json). Each file declares the ordered list of lesson slugs, prerequisites, and required assessments. Individual lessons reside in certifications/claude/lessons/<slug>/ and contain a docs/en.md explainer, implementation code, and a quiz.json. Assessments are stored in certifications/claude/assessments/<exam-code>/ and contain formal evaluation logic.

Step-by-Step Certification Workflow

1. Review the Academy Overview

Start by reading certifications/claude/README.md. This file outlines the high-level workflow, prerequisite knowledge, and the distinction between the self-paced curriculum and formal certification tracks.

2. Select Your Certification Track

Each track is defined by a JSON configuration in certifications/claude/tracks/. Available tracks include CCDV-F (Foundations), CCAR-P (Professional), CCAR-F (Foundations), and CCAO-F (Foundations). Inspect your chosen track to see the ordered lesson list:

cat certifications/claude/tracks/ccar-f.json | jq '.lessons[]'

3. Complete the Underlying Lessons

Work through the lesson slugs specified in your track file. Each lesson directory contains:

• docs/en.md – Learning objectives and theory
• code/ – Runnable implementation files (e.g., main.py)
• quiz.json – Assessment questions you must pass

Execute the lesson code to ensure it runs without errors:

python3 phases/10-agent-systems/01-agent-basics/code/main.py
python3 -m unittest discover phases/10-agent-systems/01-agent-basics/code/tests

4. Execute Certification-Specific Labs

Each certification lesson includes an Interactive Lab or Practice Lab in its outputs/ folder. Use the helper script scripts/lesson_run.py to execute the lesson and generate required artifacts:

python3 scripts/lesson_run.py \
  --lesson certifications/claude/lessons/32-architect-professional-system-capstone \
  --run-output outputs/skills/skill-arch-prof-capstone.md

The script verifies that the lesson exits with status 0 and produces the expected output files in certifications/claude/lessons/<slug>/outputs/.

5. Pass the Final Assessments

After completing all lessons in a track, run the diagnostic or final assessment located in certifications/claude/assessments/<exam-code>/:

python3 scripts/lesson_run.py \
  --assessment certifications/claude/assessments/ccar-f/assessment.json

The assessment runner validates your answers against expected indices and generates a result file upon passing.

6. Verify and Audit Your Completion

Store all generated JSON result files in your personal learning-artifacts/ directory. Run the verification script to confirm all required artifacts are present:

python3 scripts/audit_certifications.py --track ccar-f

When audit_certifications.py completes without errors, it produces a verification report confirming you have earned the Claude certification for that track. This report serves as shareable proof of competence.

Key Automation Scripts

The certification process relies on two primary automation tools located in the scripts/ directory:

• scripts/lesson_run.py – Executes individual lessons, runs labs, and processes assessment JSON files to validate technical implementation.
• scripts/audit_certifications.py – Performs a final validation check against track requirements, verifies the presence of all learning artifacts, and generates the official certification report.

Available Certification Tracks

According to the source code structure, the academy currently defines multiple specialization routes:

• CCAR-F – Claude Certified Architect Foundations
• CCAR-P – Claude Certified Architect Professional
• CCDV-F – Claude Certified Developer Foundations
• CCAO-F – Claude Certified AI Operator Foundations

Each track JSON file in certifications/claude/tracks/ specifies the exact sequence of lessons and the specific assessment files required for credentialing.

Summary

• The Claude Certification Academy is located at certifications/claude/ in the ai-engineering-from-scratch repository.
• Track definitions in certifications/claude/tracks/ JSON files dictate the required lesson order and assessments.
• Complete lessons by reading docs/en.md, implementing code in code/, and passing quiz.json validations.
• Use scripts/lesson_run.py to execute labs and assessments, ensuring they exit with status 0.
• Run scripts/audit_certifications.py --track <id> to verify completion artifacts in learning-artifacts/ and generate the final certification report.

Frequently Asked Questions

What are the prerequisites for the Claude Certification Academy?

The academy assumes you have completed the corresponding prerequisite lessons in the phases/ directory or possess equivalent knowledge in AI engineering fundamentals. The certifications/claude/README.md file explicitly lists which foundational concepts (e.g., agent systems, prompt engineering) you should master before attempting certification tracks.

How long does it take to complete a certification track?

Completion time varies by track complexity. The CCAR-F (Architect Foundations) track requires completing multiple capstone projects and assessments, typically requiring 20-40 hours of focused work. Self-paced learners progress through lesson slugs sequentially, with each lesson requiring both study time and hands-on lab execution via scripts/lesson_run.py.

Can I attempt the assessments multiple times?

Yes. The assessment runner in scripts/lesson_run.py validates answers against expected indices but does not impose attempt limits. You can re-run the assessment JSON files in certifications/claude/assessments/<exam-code>/ until you achieve a passing score, at which point you generate the result artifact for your learning-artifacts/ directory.

Where are my certification artifacts stored?

The scripts/audit_certifications.py script expects proof of completion files in a learning-artifacts/ directory at the repository root. These artifacts include JSON result files generated by successful assessment runs. The audit script checks this directory against the requirements specified in your chosen track's JSON definition to issue final verification.

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