Interview Question Generator

Ten mock interview questions for a specific role and seniority, mixing behavioural, technical, situational and culture questions, each with answer tips you can open one at a time. The same role and level always produce the same set, which makes it usable as a repeatable practice drill rather than a slot machine.

Practice That Can Be Repeated

Random question generators are close to useless for practice, because you cannot tell whether you improved or simply drew easier questions. This one seeds its selection from the role, the level and the question type, so regenerating the same combination returns the same ten questions in the same order. That single property turns it from a novelty into a drill you can run twice in a week and compare.

• The role you type is classified by keyword into one of four buckets — engineering, product, sales, or a general set — and the technical and situational questions are drawn from that bucket. Typing "backend engineer" and "platform developer" reaches the same pool, because both match the same keyword pattern.
• Four question families are represented in every run: behavioural questions, technical questions, situational questions and culture-fit questions. The mix is what stops the set from being either a coding quiz or a personality questionnaire.
• Seniority has four settings — Junior with 0 to 2 years, Mid with 3 to 5, Senior with 6 to 9, and Lead / Staff for 10 and over — and the difficulty of the drawn questions follows the selection. There is also an industry selector covering sectors such as SaaS and tech, e-commerce and finance, which shifts the framing rather than the subject matter.
• Twenty-one question templates are held in the page, merged as role-specific plus general for each family, then ordered deterministically. A ten-question set is drawn from that ordering.
• Each question expands in place to show answer tips rather than the answer. The distinction matters: a tip tells you what a strong response covers, so you still have to construct it.
• Any question can be copied on its own, which is what you want when you are working through a set across several days rather than reading it in one sitting.

Running a Session That Actually Changes Anything

1. Enter the exact job title from the advertisement you are preparing for, not a generic description. The role bucket is decided by keyword, so the title is the input that decides which technical and situational questions appear.
2. Set the seniority to the level you are interviewing at, not the level you feel. A Senior set answered with Junior-length stories is the most common way a prepared candidate underperforms.
3. Choose the industry if you have one, then generate and read the ten questions without expanding the tips.
4. Answer each question out loud against a timer before opening its tip. This is the entire value of the exercise — reading tips first converts a rehearsal into a reading task.
5. Open the tip afterwards and note what you missed rather than what you got right. The gaps are what to prepare; the successes need no further work.
6. Regenerate the same role and level a few days later. Because the seed is deterministic, you get the same ten questions and can measure whether the gaps closed.

Four Kinds of Question, Four Kinds of Failure

Alternative	Notes
Behavioural — the STAR family	Asks for a past event and is lost by candidates who describe what they would generally do instead of what they did once. The failure mode is drift from specific to general, and the fix is having three or four stories with verifiable detail ready to reuse.
Technical	Asks for a method or a mechanism and is lost by over-answering — covering six options when the question asked for one, or by the reverse, naming a framework without describing how it behaves. The failure mode is depth mismatch rather than ignorance.
Situational — hypothetical	Presents a scenario that has not happened and is lost by treating it as behavioural, reaching for a past example when the interviewer wants reasoning. The failure mode is category confusion, and it is easy to spot once you know the question is hypothetical.
Culture fit	Asks how you work and is lost by answers that would fit anywhere. The failure mode is blandness, and unlike the other three it cannot be fixed with preparation alone — it needs a genuine position on how you like to work.

Two Roles, Two Different Sets

The role bucket is the strongest lever, and seeing two titles side by side shows how much it changes.

• Enter Product Manager at the Senior level and the classifier lands on the product bucket. The technical family draws from product-oriented templates — prioritisation under constraint, metric definition, handling a disagreement with engineering — rather than from implementation questions.
• Enter Sales Manager and the bucket becomes sales, so the situational questions turn on objection handling, pipeline judgement and negotiation framing. The behavioural questions stay broadly the same shape, because a STAR question about conflict works in both roles.
• Enter a title containing no recognised keyword, such as Operations Coordinator, and the general bucket supplies the set. This is not a failure state: the general templates cover the questions that every interviewer asks, and the answer tips are the same ones that apply to the specialised sets.
• Change only the seniority on a fixed role and the questions change while the families stay the same. Junior sets lean toward prompts asking you to describe a process you followed, Lead sets toward prompts asking how you would change one.
• Because all of this is seeded rather than random, two people preparing for the same role at the same level get identical sets — which is what makes a set discussable in a study group rather than a private draw.

Where the Questions Come From

The set is assembled from templates compiled into the page and ordered by a seed derived from your inputs. There is no model in the default path, which is why the output is instant, reproducible, and free of the confident waffle that question-generating models produce. The limit of that approach is coverage: twenty-one templates across four families and four role buckets means a very specialised role will draw a set that is competent rather than tailored, and the technical questions will test general method rather than deep domain specifics.

• Answer tips are written per template, so a tip is genuinely specific to its question rather than a generic STAR reminder.
• The deterministic ordering means the tool cannot surprise you twice, which is a feature for drilling and a limitation if you want variety — changing the seniority setting is the fastest way to a different set.
• Nothing is stored between sessions and no account is needed. Refresh and the set is gone unless you copied it.
• This tool's provider list is shorter than most on the site, offering DeepSeek, OpenAI, Anthropic, Google and a Custom endpoint. The built-in path does not need any of them.
• If you do connect a provider, the role title and seniority are what get sent, so keep the input to the job description rather than pasting details about a specific employer's internal structure.

Questions About Using the Set

Why do I get the same questions every time for the same role?

Because the ordering is seeded from the role, the seniority and the question family rather than drawn at random. For interview practice that is the point — you can run the set on Monday, work on the gaps, and run the identical set on Friday to see whether the answers improved. If you want a different set, change the seniority setting, which reseeds the draw.

How are the questions matched to the role?

By keyword. Titles containing engineering or developer terms such as backend, frontend, fullstack, SRE, DevOps or architect land in the engineering bucket; product and PM titles in the product bucket; sales and account-executive terms in the sales bucket; anything else falls to a general set. That means a specialised title can land in the general bucket, which is a reasonable outcome rather than an error — the general set is the set every interviewer draws from.

Are the answer tips the answers?

No, and treating them as answers is the fastest way to waste the tool. Each tip outlines what a strong response covers — the structure, the elements a listener expects to hear, the thing that tends to be omitted. You still have to build the answer from your own experience, which is what the interviewer is assessing.

Is anything about the role I type sent to a server?

Not in the default mode. The classification and the ordering both happen in the page against compiled templates, so the role title never leaves the tab and the tool works with the network disconnected. If you connect an AI provider, the role and level become part of the request — worth knowing before typing an internal job title that has not been announced.

Last verified: October 3, 2026

Launch Interview Question Generator

All Tools · AI Tools · About · How We Test · Contact · Privacy

Maintained by Huang Bin · Last updated October 3, 2026