In this article
  1. Start here: the learning path
  2. What are the stages of a typical interview process?
  3. What happens in an interview loop?
  4. What is a technical interview?
  5. How do AI-lab interview processes differ?
  6. How do I prepare for a technical interview?
  7. Interview rounds by type
  8. Questions interviewers ask at each stage
  9. All 60 guides in Interview loops and rounds
  10. Follow-up questions
  11. Sources

Start here: the learning path

Interview process
An interview process is the ordered sequence of applications, assessments, conversations and decisions that a company runs to choose one candidate for a role, with each stage designed to remove candidates who lack a specific signal.
Technical interview
A technical interview is a conversation or exercise in which a candidate solves or explains problems drawn from the skills of the job, such as coding, system design, data work or machine learning, instead of only describing past experience.
Screening call
A screening call is a short first conversation, usually by phone or video, in which a recruiter or interviewer checks fit, motivation and basic skill before the company invests in a full loop.

Follow this order if you are early in a search. Each step builds on the one before it.

  1. Read the stages on this page so you know what each stage is trying to learn about you.
  2. Prepare a two-minute answer to "tell me about yourself", since many screening calls begin there.
  3. Prepare your stories for behavioral questions.
  4. Prepare the technical rounds your role uses: coding interviews, system design, machine learning (ML) system design or ML coding.
  5. Learn the loop of the company you are targeting, starting with OpenAI interview questions and the Anthropic interview process.
  6. Prepare the close: questions to ask the interviewer and your salary negotiation.
The stages of a typical interview processA left-to-right flow of five stages: online application, assessment, screening call, interview loop and decision. The first four are filters and the loop produces the written feedback that the decision combines.Applicationrole matchAssessmentskill at scaleScreeningcallfit, motivationInterviewloopseveralsessionsDecisioncombinedfeedbackDiagram: Interview Clozo, CC BY 4.0 ·interview.clozo.ai/blog/interview-loops/
Most processes run in this order, and each stage filters for something different.CC BY 4.0 — source: What is the interview process like? Stages, rounds and who scores what
Text description of the diagram
  • Amazon lists four stages: online application, assessments, phone screening and interview loop, and says the process differs from role to role.
  • The decision follows the loop and combines the interviewers' written feedback.
  1. Code review interview questions: what to check, feedback and AI-written code
  2. Meta engineering manager interview: rounds and how to prepare
  3. What is system design for managers, and how do you answer it?
  4. How do AI-engineering take-homes work in 2026?
  5. How do interviewers probe time and space complexity?
  6. What is the staff follow-up in a coding interview?
  7. What is the AI-allowed build exercise, and how do you handle it?
  8. What is the project deep dive, and how do you answer it?
  9. What is the refactor take-home, and how do you do it well?
  10. What do forward deployed and applied AI engineer rounds ask?
  11. What is the project retrospective round, and how do you answer it?
  12. What is the coding interview for engineering managers?
  13. What is the eval take-home, and how do you answer it?
  14. How interview loops change from senior to staff to principal: more design rounds
  15. What is an AI-assisted coding interview, and how is it graded?
  16. Amazon applied scientist interview process: rounds and preparation
  17. OpenAI interview questions and process: what to expect and how to prepare
  18. ML system design interview: how the round works and how to run it
  19. How do you do a machine-coding or low-level design take-home, step by step?
  20. What is the RAG take-home, and how do you do it well?
  21. What is the agent take-home, and how do you pass it?
  22. The classical-ML take-home: what it is and how to do it well
  23. The extraction take-home: how to build and defend it
  24. The data/analytics agent take-home, step by step
  25. Multi-agent and context-engineering take-homes: memory per agent, step by step
  26. The MLOps take-home: what it is and how to do it well
  27. How do you build a mini workflow engine (mini-LangGraph)?
  28. How to prepare for a phone interview
  29. What are non-leetcode functional rounds, and how do you answer them?
  30. What is the manager role-play round, and how do you handle it?
  31. What are EM and staff presentation rounds?
  32. Machine learning breadth interview questions and answers
  33. What project are you most proud of?
  34. Python interview questions on types, hashing, arguments, generators and threads
  35. How have tech interviews changed in 2025–26?
  36. Is this now the standard for all levels including E6?
  37. What to write in a writing sample or tech spec take-home
  38. Why did you accept that code? The design, test and simplicity checks
  39. Tell me about a management decision you'd reverse if you could
  40. How do you decide what to use versus reject from the AI's output?
  41. How do you handle performance issues or challenges with team members?
  42. How do you approach one-on-ones with your team, and what do you talk about?
  43. Prioritising under pressure as an IC: how to answer
  44. Delivery under pressure: cut scope, not quality, and warn stakeholders early
  45. What's your management philosophy? Two principles, each tied to a behaviour
  46. Why do you want to work here? How to answer it, with examples

What are the stages of a typical interview process?

A typical interview process has an application, an assessment, a screening call, a loop and a decision. Amazon lists four stages: online application, assessments, phone screening and interview loop, and says the process differs from role to role (Amazon, read October 2026). The decision follows the loop and is made by combining the interviewers' written feedback.

In many processes each stage removes candidates for a different reason. The application checks that your experience matches the role. The assessment, when there is one, checks a skill at scale. The screening call checks motivation and communication. The loop checks depth in several areas by independent people. Companies change the order: some put the technical exercise before the first call, and some skip the assessment for experienced hires. Ask the recruiter for the exact stages in your first conversation, because the plan differs by team and level.

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 →

What happens in an interview loop?

In an interview loop you meet several people one at a time, and each assesses a different aspect of your skills and experience. Amazon's loop page describes meeting individually with current employees, with each person assessing different aspects, and names leadership principles and behavioral-based questions answered with the situation, task, action, result (STAR) method (Amazon loop page, read October 2026).

Many companies standardise the loop. Google's guide describes structured interviewing as asking every candidate for the same job the same questions, graded on the same scale with rubrics that define outstanding, solid, borderline and poor answers, plus interviewer training and calibration (Google re:Work, updated March 2026). For you, this means each session has one purpose and a scoring scale. Each interviewer scores their own session against its rubric, and the decision weighs all the scores together, so prepare every round instead of counting on one strong session to carry the loop.

What is a technical interview?

A technical interview tests the skills the job uses by having you solve or explain problems live. For an engineering role that is usually coding, system design or both. Anthropic's careers page describes live coding exercises in tools such as Colab and CodeSignal for technical roles, says looking things up is allowed, and expects candidates to be comfortable with basic syntax and standard libraries (Anthropic, read October 2026).

A technical interview is not only about the final answer. The interviewer listens to how you clarify the problem, how you reason aloud, how you test your own work and how you respond to a hint. For data and machine-learning roles the same shape applies with different material: data engineering, machine learning operations (MLOps) and ML coding each have their own rounds.

How do AI-lab interview processes differ?

AI-lab processes follow the same stages but add explicit rules about AI use and often a longer work sample. Anthropic's candidate guidance, last updated July 2025, asks applicants to write first drafts of application materials themselves and says live interviews are done without AI assistance unless the company indicates otherwise (Anthropic, July 2025). Its January 2026 engineering post describes a take-home for performance engineers that explicitly allows AI use, because longer-horizon problems are harder for AI to solve completely (Anthropic, January 2026).

The practical lesson is to read the instructions for every stage. The rule can change between a live round and a take-home at the same company. The Anthropic interview process page covers one lab in detail, and OpenAI interview questions covers another.

How do I prepare for a technical interview?

Prepare for a technical interview by practising the skill under the conditions of the round: a time limit, no copy-paste of a finished solution, and speaking your reasoning aloud. Start from the job description. List the two or three round types the loop will contain, then give each a block of practice in this order: fundamentals, timed problems, then explaining your solution to another person.

Use the round map below to match each round to its preparation. After each practice session write down one thing you did well and one you will change, and re-test the change the next day. Rehearse the answers out loud, because a technical interview rewards clear speech as much as correct code.

Interview rounds by type

Coding and ML coding rounds

A coding round asks you to write working code for a stated problem while talking through your approach. The interviewer scores problem understanding, correctness, complexity and testing. Prepare with coding interviews for general algorithms and ML coding for implementations such as attention or a training loop.

System design and ML system design rounds

A design round asks you to build a system on a whiteboard or document and defend your trade-offs. The interviewer scores requirements gathering, a workable architecture, capacity reasoning and communication. Prepare by designing two or three common systems end to end under a time limit and saying each trade-off aloud.

Take-home assignments

A take-home assignment is a work sample you complete alone, so it tests what you ship when no one watches. Reviewers read your code, your write-up and the trade-offs you explain. Anthropic's post notes candidates work in their own editor and that more than 1,000 people had taken its performance take-home (Anthropic, January 2026). Prepare with the coding and design guides above, under a time limit.

Behavioral and leadership rounds

A behavioral round asks you to describe past situations and a leadership round asks what you decided and how you made it hold. Both reward a stated scope and a result. Prepare with behavioral interviews and leadership interviews, and for manager roles read engineering management interview questions.

Company loops and level changes

A company loop is the specific mix of rounds one employer uses, and it changes by level. Ask your recruiter which round types your loop contains and prepare each one from the matching guide.

Round types, what they test and how to practiseA table of five round types (coding, design, take-home, behavioral, leadership) with what each tests and how to practise it. Coding tests correctness, design tests trade-offs, and take-homes test shipped work.RoundTestsPractise byCodingCorrect, testedcodeTimed problems,aloudDesignTrade-offs,capacityWhole designs,timedTake-homeWork you shipaloneTime-boxedsampleBehavioralPast actionsSTAR storiesLeadershipDecisions atscopeDecisions andresultsDiagram: Interview Clozo, CC BY 4.0 ·interview.clozo.ai/blog/interview-loops/
Match the round to what it tests, then prepare the skill and not the format.CC BY 4.0 — source: What is the interview process like? Stages, rounds and who scores what
Text description of the diagram
  • Coding rounds test correctness, complexity and testing; design rounds test requirements, architecture and trade-offs.
  • A take-home assignment tests what you ship alone; a behavioral round asks about past situations; a leadership round asks what you decided.

Questions interviewers ask at each stage

Question lists help once you know the round. The AI engineer interview questions page collects answered questions across AI engineering rounds, including a section on agents and tools.

Follow-up questions

What does a technical interview involve?

A technical interview has you solve or explain problems from the job's skills, and Anthropic's careers page describes live coding exercises where looking things up is allowed (Anthropic, read October 2026). See coding interviews for the round in detail.

What comes after a technical interview?

Interviewers record their feedback so the people deciding can review it, and the decision is made against qualifications set in advance (Google re:Work, March 2026). Prepare your salary negotiation before an offer arrives.

How long is a technical interview usually?

It varies by company and round. Ask the recruiter for the length of each session, then practise to that limit with coding interviews.

What are the three stages of an interview?

There is no single standard. Amazon lists four stages and says they differ by role (Amazon, read October 2026); many processes reduce to a screen, a deeper assessment and a final loop; the stages are described in the stages section above and for one lab in the Anthropic interview process.

Key takeaways
  • An interview process is a series of filters, and each stage tests something different.
  • A loop is several separate sessions, so each session is scored on its own.
  • Read the instructions of every stage, especially the rules on AI use.
  • Prepare the skill of each round under the round's conditions, aloud.

Sources

  1. Interviewing at Amazon — Amazon, 24 Oct 2026
  2. Amazon interview loop — Amazon, 24 Oct 2026
  3. Anthropic careers: interview format — Anthropic, 24 Oct 2026
  4. Anthropic guidance on AI use in applications (last updated July 2025) — Anthropic, 10 Jul 2025
  5. Designing AI-resistant technical evaluations — Anthropic (Tristan Hume), 21 Jan 2026
  6. A guide to structured interviewing for better hiring practices (updated March 2026) — Google re:Work, 1 Mar 2026