Module 10 — Certification Exam¶
Duration: ~1 hour (90-minute hard cap) · Status: Required to earn certificate
This is the final module. Passing it earns the Asfela AI Security Engineering — Professional certificate.
What this module contains¶
| File | Purpose |
|---|---|
L10.1-exam-instructions-and-format.md |
Rules, format, time limit, pass mark, retake policy |
L10.2-exam-item-bank.md |
50 questions covering M0–M9 |
L10.3-answer-key-and-explanations.md |
Correct answers + rationale + lesson reference per question |
L10.4-scoring-and-retake-policy.md |
How to score yourself; retake conditions; certificate issuance |
99-course-completion.md |
Course-wide closing — what you learned, what's next, alumni paths |
Order to use this module¶
- Read L10.1 (instructions, ~5 min).
- Take the exam in L10.2 under timed conditions. Do not look at L10.3 first.
- Self-score against L10.3 (or have the LMS auto-grade — answer key included for both modes).
- Read L10.4 for the pass/fail decision, retake policy, and certificate issuance.
- Read 99-course-completion.md — closing module, alumni paths, next steps.
Exam format at a glance¶
- 50 questions total.
- 40 single-answer multiple choice (4 options each, pick one).
- 10 scenario-based (a short situation followed by a single best-answer question).
- Pass mark: 75% (38 of 50 correct).
- Time limit: 90 minutes (target 60 min; 30 min buffer).
- Closed book in spirit (you may consult the course modules for concept reminders but not for direct question-answer lookup — these are conceptual questions, not trivia).
- Attempts: 2 attempts permitted; 7-day cool-down between attempts.
Coverage map¶
| Module | Topic | # of questions |
|---|---|---|
| M0 | Orientation, course logistics | 1 |
| M1 | AI/ML foundations | 4 |
| M2 | AI security foundations (threat modeling, ATLAS, OWASP, frameworks) | 6 |
| M3 | Prompt injection & LLM app attacks | 9 |
| M4 | Data poisoning, backdoors, supply chain | 6 |
| M5 | Extraction, inversion, membership inference | 5 |
| M6 | Adversarial examples & evasion | 4 |
| M7 | MLSecOps & defenses | 7 |
| M8 | Governance, risk, compliance | 6 |
| Cross-module | Scenario items spanning multiple modules | 2 |
| Total | 50 |
Why these weights¶
The question count per module roughly tracks the module's depth and the frequency of the topic in real AI security engineering work. Prompt injection (M3) and MLSecOps defenses (M7) get the most weight because they are the most operationally relevant. Governance (M8) is heavily weighted because regulators reward proof of understanding, and your employer needs that. M0 is a single sanity question.
After the exam¶
Pass → certificate issued through Thinkific (or whichever LMS hosts the course). Your name + course name + cert ID + date → PDF + verifiable URL. Add to LinkedIn, resume, GitHub.
Fail → review L10.3 explanations, revisit weak modules, retake after 7 days.
Either way: read 99-course-completion.md for the closing module.