Skip to content
AI Security

Interview question · How LLM Systems Work

Describe a RAG pipeline as a data flow and mark every point an attacker can influence.

An easy question with a lot of surface area. The scored answer walks the pipeline left to right, names the attacker's lever at each stage, and ties each to one control.

easyEssentialNewUpdated Nov 2026

An easy question with a lot of surface area. The scored answer walks the pipeline left to right, names the attacker's lever at each stage, and ties each to one control.

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 5 follow-ups like an interviewer would

A fictional makerspace assistant searches manuals, member notes and restricted staff records. It uses an external reranker, generates cited answers and stores a tenant-wide answer cache.

  1. What if staff text reaches the external reranker before the user-access filter?
  2. What if a permitted member note contradicts the official equipment manual?
  3. What if the index denies a revoked source but its old answer still comes from cache?
  4. What if each tenant has a separate index but the caller chooses the index name?
  5. What would your review deliver?

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 trackHow does function calling work, and where does an agent actually get its authority?Next in this trackWhat are embeddings, and why do they matter for security filters and for data leakage?Next in this trackIf the model is forced to return valid JSON matching a schema, is its output safe to use?
DISCUSSION · 0

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