In this article
What is an interview preparation guide?
- Interview preparation guide
- An interview preparation guide is a page that teaches one topic an interviewer tests, from first principles to an answer you can say aloud, and the follow-up questions that come next.
- Hub guide
- A hub guide is the overview page of one topic, which defines its core terms, sets a reading order and links every other page about that topic.
- Question index
- A question index is a list of at least fifteen distinct interview questions on one topic, each with a short answer, the reason it is asked and a link to the full explanation.
- Model answer
- A model answer is a worked example of a strong spoken answer, short enough to say in about a minute, that shows the structure rather than words to memorise.
A rubric is the scoring table interviewers use; single-question guides include one so you can grade your own answer.
Start here: the learning path
Choose the line below that matches your situation, then open its first guide. Each guide has its own path to follow.
- I do not yet know what rounds I face. Read how interview loops work: the stages from application to decision, and what each stage tests.
- I am preparing for an AI or machine-learning (ML) engineering loop. Read the AI engineer question list, then the guides on transformers, large language model (LLM) inference and AI agents.
- I have a design round. Read ML system design and system design.
- I have a coding round. Read coding interview questions and the ML coding guide.
- I have a behavioural or leadership round. Read behavioral interviews, then leadership interviews.
- I have an offer. Read how to negotiate salary.
Knowing it is half the job. Saying it is the other half.
Interview Clozo is an AI interview copilot for real interview rounds: it hears the question and puts a strong answer on your screen, and you say it in your own words. See the features or how it works.
Start Free Trial →How is each guide built?
Structured interviews are scored against a rubric, and the guides are built so you can practise against one. In a structured interview, candidates get the same questions and the answers are rated on one scale against predetermined criteria; Google describes the method as "using the uniform methods to assess candidates applying for the same job" (Google re:Work, updated March 2026). Not every interview is that structured; a review of about twenty years of research covers how structure is defined, how rating scales are built and how probing questions work (Levashina, Hartwell, Morgeson and Campion, 2014).
Each topic guide opens with the answer and defines every term; concept and problem guides then explain the mechanism step by step and work a numeric example, and single-question guides include a Staff-level answer and a rubric. Hub guides, like this page, give the overview and the reading order. Practice should copy the real condition: in a randomised controlled experiment with 48 computer science students, performance fell by more than half when an interviewer watched a candidate solve a problem, compared with solving it privately (Behroozi, Shirolkar, Barik and Parnin, 2020). Say your answers aloud.
How is the library organised?
The library has six families, and each guide is filed under one of them. Each family has one or more hub guides that define its terms and link its pages.
Text description of the diagram
- The library has six families, and each guide is filed under one of them.
- Each family has a hub guide that defines its terms and links its pages.
AI and machine-learning concepts
The concept family explains how modern ML models work and how they are served. It holds the guides on transformers, deep learning, classical machine learning, LLM inference, AI agents, embeddings and vector search, prompting and machine learning operations (MLOps). Pages inside them include the key-value (KV) cache, FlashAttention, byte-pair encoding, gradient descent, activation functions, the Model Context Protocol and the confusion matrix.
Maths, probability and statistics
The maths family covers the parts of mathematics that interviews use. Its hub is statistics interview questions, and its pages include the Gaussian distribution, linear algebra for machine learning, entropy, cross-entropy and Kullback-Leibler (KL) divergence and cross-entropy loss. Read it when a loss function, a metric or an experiment question appears in a loop. Start with distributions: linear algebra does not depend on them, and sampling, tests and information theory all build on them. The family teaches the ideas that appear again and again in model-evaluation and experiment-design questions, and it links out to the model pages that use them.
System design and coding
The design and coding family covers the rounds in which you build something live. It has the hubs for ML system design, system design, ML coding and coding interview questions. The design guides teach a repeatable structure for open-ended problems; the coding guides teach the pattern map and a six-step routine. Read the design hub first if your loop has an open-ended round, and the coding hub first if it has timed problems. The ML versions of both add the parts that machine-learning roles are asked, such as writing a model component from scratch or sizing the serving system for a model.
Behavioural, leadership and management
The people family covers rounds about how you work. Its hubs are behavioral interviews, leadership interviews, engineering management and product management. Its single-question pages, such as Tell me about yourself and questions to ask the interviewer, answer the questions most loops contain. Start with the hub that matches the role you are applying for, then read the single-question pages in the order you expect the questions to arrive: the opening question first, the questions you ask last.
The interview process and employer loops
The process family covers what the stages are and what to expect from them. It holds how interview loops work, the Anthropic interview process and OpenAI interview questions and process, which walks through that employer's five stages. It is the place to start when the rounds are unknown. Each process page describes stages and what each one tests, using public sources where they exist and plain wording where they do not, so you can plan the weeks before the first call.
Careers, pay and adjacent roles
The careers family covers pay and roles beside software and ML engineering. It holds how to negotiate salary, data engineering and finance and investment banking. Read the salary guide before the offer arrives, not after, because the preparation steps come first.
More from the blog
The blog also carries research and buyer's guides about AI interview copilots, kept beside the interview guides. Start with the AI interview copilot buyer's guide, compare tools in the best AI interview copilots of 2026 and AI interview copilot pricing, read the AI interview copilot latency report and coding-round tools compared with an interview copilot, and see how to use AI in live interviews ethically, why real-time reasoning beats memorization and the investor questions founders face.
Questions interviewers ask, by topic
The question indexes list recurring questions in short form, each linking to the full explanation. They are the AI engineer question list, agentic AI interview questions and prompt engineering interview questions. Employer pages, such as the OpenAI interview process guide, describe the stages rather than list questions. The glossary collects the terms the guides define, and the diagram library shows the mechanisms visually.
Follow-up questions
What's the best way to prepare for a job interview?
Find out the stages first, then practise each one aloud. How interview loops work lists the stages from application to decision; then open the hub for the round you face, such as behavioral interviews or the AI engineer question list, and follow its path in order.
What is the first thing you should do before an interview?
Learn what the rounds are and what each tests. The stage list in how interview loops work answers that, and it decides which family of guides to open first.
How do I calm my nerves before an interview?
Rehearse under the real condition: aloud and timed. The observer study above did not test practice, but it found that being watched cut performance by more than half, so rehearse at least some answers in front of another person (Behroozi, Shirolkar, Barik and Parnin, 2020). Keep one opener ready, such as Tell me about yourself, and read the behavioral interviews hub for the story format.
What questions should I ask the interviewer?
Prepare three that show how you think about the role, the team and success in the first months. Questions to ask the interviewer lists them by round and explains what each signals.
- Pick the guide that matches your next round; each has a path.
- Single-question guides include a rubric, because structured interviews are scored against one.
- Practise aloud and with an observer; being watched is the hard condition.
- Use question lists to find gaps and guides to close them.
- The glossary collects the terms the guides define, each in one place.
Sources
- A guide to structured interviewing for better hiring practices — Google re:Work, 1 Mar 2026
- The structured employment interview: Narrative and Quantitative Review of the Research Literature (Personnel Psychology 67(1), 241-293) — Wiley (Levashina, Hartwell, Morgeson, Campion), 7 Oct 2013
- Does Stress Impact Technical Interview Performance? (Behroozi, Shirolkar, Barik, Parnin, 2020) — ACM, 8 Nov 2020
