How we write and check our questions

The short answer

We draft questions and guides with the help of AI, from the kinds of question UK graduate employers use and from frameworks they publish. The Deeplink Interview team then reads and edits them before they’re added. Each question gets a public guide written for it, and private marking notes that stay on our server. Our build checks the links, the notes and the coding challenges’ answers automatically, and every page shows when it was last reviewed.

Where the questions come from

• The types of question graduate interviews use: background, motivation, competency (“Tell me about a time…”), strengths, values, situational judgement, technical questions for data roles, estimates, cases and presentations. Our interview questions library has a guide to each type.
• Frameworks that are published: for example the Civil Service Success Profiles and the NHS values. We explain them from the published material.
• What data teams ask about: SQL, cleaning and joining data, statistics, charts, pipelines and data modelling, explained out loud as you would to an interviewer.

We don’t use leaked or confidential questions, we don’t work with or speak for any employer, and we never say that a particular employer asks a particular question or how it scores answers.

How a question is written

1. Drafted with the help of AI. A draft question, its type and its timings are written from the question type and the published frameworks above.
2. Read and edited by a person. The Deeplink Interview team checks each question before it’s added to the bank: that it’s clear, fair, and answerable by a graduate without inside knowledge; that it isn’t a trick question; and that it’s in plain English. Questions never ask about protected characteristics, such as age, health or religion.
3. A public guide, written fresh. Each question has its own page: what it’s really asking, how to structure an answer, and an example outline. The outlines are invented to show the shape of a strong answer. None is a real person’s answer.
4. Private marking notes. Each question also has short notes on what a strong answer includes, which our AI marking uses to sharpen its feedback. They stay on our server and are written separately from the public guide. How marking works explains the rest.

What’s checked automatically

Every time the site is built and tested:

• Every question in the bank must have marking notes, or a test fails.
• A question page without a written guide is kept out of search results and out of our sitemap until its guide is written.
• Every link from a guide, study plan or role page to a question, set or coding challenge must point to one that exists, or the build stops.
• Coding challenges are run. Each one’s model answer is run against its hidden tests, and each common slip it warns about is checked not to pass. The expected results are worked out from the model answers, and a test fails if they’re out of date.

Coding challenges

Our SQL, pandas and Python challenges are original, set in invented UK organisations with made-up data. They run in your browser and are marked by comparing results, not code, so any correct answer passes. No AI marks them. Try one.

Reviews and dates

Each question page and each interview guide shows when it was last reviewed. We review a page again when a framework it explains changes, or when someone tells us something is wrong or out of date.

Reviewed by the Deeplink Interview team.

Something wrong?

If a question, guide or challenge is unclear, out of date or wrong, email us at coaching@deeplink.uk with the page’s address. We’ll look into it, correct anything that’s wrong and update the page’s date.

What this isn’t

Our questions are for practice. They aren’t any employer’s real questions, and nothing here, or in our feedback, predicts a hiring decision.