AI security questions.
Reason through the decisions.
Choose a topic, attempt a question and compare your reasoning with the worked answer. Use the follow-ups to change an assumption, then check your explanation against the scoring criteria.
- 01
You are handed a system you have never seen. How do you find its trust boundaries, and why start there?
The first ten minutes of any design review decide its quality. This answer shows the sketch, the one question that finds most bugs, and why controls come last.
Answer available - 02
How is securing an LLM application different from ordinary application security, and what stays exactly the same?
Candidates who say 'everything is new' and candidates who say 'nothing is new' both lose this one. The answer that scores names exactly three differences and keeps the rest of the toolkit.
Answer available - 03
Walk me through threat modeling with data-flow diagrams and STRIDE. When do you do it, and what do you produce?
Reciting the six letters earns little. What scores is a diagram drawn aloud, one boundary walked in depth, and a list of threats with owners.
Answer available - 04
Threat-model a customer-support chatbot that answers from a knowledge base and can look up orders.
The order lookup is the whole interview. Candidates who spend ten minutes on jailbreaks and two on authorization have it backwards.
Answer available - 05
Explain least privilege and defence in depth with an example where one layer failing does not cause a breach.
Definitions take ten seconds. The interviewer wants a worked failure: which layer broke, which one held, and why they did not break together.
Answer available - 06
What is the difference between a threat, a vulnerability and a risk, and how do you prioritise findings?
Three words people use interchangeably, and a prioritisation question that exposes whether you have ever had a backlog of thirty findings.
Free with an account - 07
A product team wants to ship an LLM feature next week, and the security review is not done. What do you do?
Neither 'block it' nor 'ship it' scores. The interviewer is watching whether you can find the one risk that matters in a day and offer a path to launch.
Free with an account - 08
Explain an AI security risk to a CISO: the impact, the options, the residual risk and the decision you need.
Five minutes, no jargon, one decision. Candidates who explain how attention works have already lost the room.
Free with an account - 09
Walk through the incident response lifecycle using a leaked API key as the example.
Most candidates start by deleting the commit. The key was copied the minute it went public; the clock starts with revocation.
Free with an account - 10
Threat-model a password-reset flow.
A reset flow is a second login page with weaker checks. Interviewers use it because every step has a classic bug, and the strong answer finds them in order.
Free with an account - 11
How do attack trees and kill chains help in threat modelling, and where do they mislead?
Two classic tools that structure attacker thinking. The scored answer uses each for what it is good at and names the blind spots before an interviewer does.
Premium answer - 12
Why does STRIDE need adapting for ML components, and how would you adapt it?
STRIDE assumes components follow their code. A model follows its inputs. The scored answer keeps the six categories and adds what probabilistic, persuadable components need.
Premium answer - 13
What would you log for a new service, and how do you write a detection with an acceptable false-positive rate?
Detection quality is decided before the first alert, by what you logged and how you set the threshold. The scored answer derives both from the questions an investigator will ask.
Premium answer - 14
Apply data minimisation and purpose limitation to a feature that sends user data to a model provider.
Privacy principles become engineering decisions about fields, retention and contracts. The scored answer walks one feature and makes each principle concrete.
Premium answer - 15
How do you run a security design review, and how do you write findings that engineers actually act on?
Reviews fail when they arrive late, cover everything equally and produce vague findings. The scored answer scopes early, focuses on the riskiest decisions and writes findings as specific, owned changes.
Premium answer - 16
Tell me about a security risk you could not eliminate. How did you handle it?
Every senior security engineer has one. The scored answer shows you reduced the risk, made the residual visible, got the right owner to accept it, and kept watching.
Premium answer - 17
Tell me about a time you pushed back on a product team for security reasons.
Interviewers want a security engineer who can say no without becoming the department of no. The scored answer shows evidence, an alternative that met the goal, and a working relationship afterwards.
Premium answer - 18
You and an engineering team disagree on a finding's severity. How do you resolve it?
Severity disputes are usually disagreements about facts or assumptions. The scored answer finds which, settles it with evidence and the agreed scale, and has an escalation path that does not damage trust.
Premium answer - 19
Where do security activities sit in the lifecycle of an AI feature, from idea to retirement?
AI features need the usual secure development steps plus a few that most lifecycles lack. The scored answer places each activity at the stage where it is cheapest and names the gate.
Premium answer - 20
An external researcher reports a flaw in your AI feature. What is your process?
How you treat the first reporter decides whether there is a second. The scored answer acknowledges quickly, validates safely, fixes by severity, communicates honestly and handles AI-specific scope questions.
Premium answer - 21
How should an application, and an AI agent, handle secrets such as API keys and database passwords?
Leaked keys remain one of the commonest breach causes, and agents create new places to leak them. The scored answer prefers no secret at all, then short-lived ones, and keeps them out of every model context.
Premium answer - 22
You are starting an AI security programme with a fixed budget. What do you do in the first 90 days, and what do you fund first?
A leadership question. The scored answer finds out what exists, fixes the highest-impact exposures, and builds the few shared controls that make every later project cheaper.
Premium answer - 23
What metrics would you track for an AI security programme?
Good metrics change decisions. The scored answer picks a few in coverage, risk, outcomes and speed, defines each precisely, and explains what action each one triggers.
Premium answer - 24
How do you build security culture in teams you do not manage?
Security engineers rarely have authority over the teams whose code matters most. The scored answer makes the secure path the easy one, earns trust through help, and gives teams ownership.
Premium answer