Certification
Certification: Self-Critique Prompting
Prove you can use Self-critique in your own job: five written questions made for your work, scored by an AI examiner. Certificate for LinkedIn.

The exam
Take the certification exam
No multiple choice: five written questions about your own job, scored by an AI examiner against the course rubric. Pass with 70% or more and your certificate is issued at once, with a public page and an "Add to LinkedIn" button.
What the exam covers
Written questions about your own job, across the 6 chapters
No multiple choice. You tell us your job and a real task; you get five open questions written for that job, such as writing a full prompt with the framework, fixing a weak prompt and choosing the right framework. An AI examiner (DeepSeek) scores each answer against the course rubric and explains the score. Pass with 70% or more.
- 01The first draft is not the answer3:26
- 02Self-refine: draft, critique, revise2:20
- 03Chain of verification2:23
- 04Critique against principles and rubrics2:45
- 05Pre-mortem and devil's advocate2:27
- 06Building a review step into your job2:45
About this certification
This certification shows that you can use Self-refine, Chain of verification, Constitutional critique, Pre-mortem in real work, not just recite the letters. It belongs to the course Self-Critique: Self-Refine, Chain of Verification and Pre-Mortems.
How the exam works
The exam has no multiple choice. You tell us your job and one real task. You then get five written questions made for that job: write a complete prompt with the framework, fix a weak prompt, choose the right framework for a situation, and explain why a part of the framework matters. An AI examiner (DeepSeek) scores each answer from 0 to 10 against the course rubric and explains the score. Pass with 70% or more and your certificate is issued at once, with a public page and an "Add to LinkedIn" button. You can retake it with new questions.
What you will be able to do
- Explain why a first draft is not the answer
- Run a self-refine loop on any text task
- Apply chain of verification to check every claim
- Build a five to seven category rubric for reviews
- Use pre-mortems and devil's advocate to surface risks
- Attach a review prompt to every important task