Skip to content
AI Security

Interview question · Cloud, Container & AI Infrastructure Security

Threat-model the path a model takes from training checkpoint to serving node.

Weights pass through more systems than most teams draw. The scored answer maps every hop, asks what an attacker could steal or swap at each, and puts integrity checks where they cannot be skipped.

hardNewPreparation tags: Anthropic, OpenAI, Google DeepMindUpdated Nov 2026

Weights pass through more systems than most teams draw. The scored answer maps every hop, asks what an attacker could steal or swap at each, and puts integrity checks where they cannot be skipped.

Make your first attempt

What would you say before reading the answer?

  1. 01 State the assumptions your answer depends on.
  2. 02 Explain the mechanism and choose a control.
  3. 03 Name a condition that would change your decision.

The worked answer includes an approach, scoring criteria and follow-up questions. Compare the reasoning with your own.

10 answers per topic instead of 5, plus saved progress and bookmarks · costs nothingor unlock all 529 remaining answers · $25

Every worked answer in all 17 topics, every lesson in every course, the full concept curriculum, the complete practice-test bank, and the downloadable security field guides.

6 months · One payment · No auto-renewal

What if · the follow-ups in this answer

The answer works through 6 follow-ups like an interviewer would

A model goes from training checkpoint to evaluation, conversion and quantisation, registration and production serving through several systems run by different teams.

  1. What is the path?
  2. What can go wrong at each hop?
  3. How do you protect integrity?
  4. How do you protect confidentiality?
  5. What about the people and pipelines in between?
  6. What controls would you put along the path?

Each rung has a nudge, a model answer, the principle that carries over and the weak answer to avoid. Premium opens every answer and ladder.

UP NEXT ON YOUR JOURNEY
Next in this trackYou discover an infected host in your GPU fleet. What do you do?Next in this trackBuffer overflows, ASLR and DEP: why do memory safety issues still matter in ML tooling?Next in this trackDesign an inference platform's control plane so that a compromised serving node cannot replace, roll back or steal models.
DISCUSSION · 0

No comments yet. Share a correction or a different way to answer it.