How to prepare for a software engineer video interview (HireVue and one-way)

To prepare for a software engineer video interview, treat it as a recorded behavioral screen plus a verbal walkthrough, not a coding test. Expect 3 to 5 questions, a short prep window, and 60 to 180 seconds to record each answer. Practice talking through your reasoning out loud, name a real project, and lead with your point.

The native term for this is a one-way interview, also called an on-demand or pre-recorded interview. If your invite came from HireVue, Spark Hire, Willo, or VidCruiter, that is the same format. This page is the full prep flow. For the exact question list and three worked model answers, pair it with one-way interview questions for software engineers.

First, what this round actually tests

Here is the part that calms most engineers down. The recorded round is rarely where a company tests whether you can code. Real coding usually lives in a take-home, a live pairing session, or an onsite. Recorded whiteboard tools exist, but recruiters who have tried them tend to find them awkward, so most teams keep hard coding out of the recorded stage. That is good news for your prep: you can focus on talking clearly rather than writing under a timer.

So the recorded round is usually checking three things:

• Communication. Can you explain a technical idea to someone who is not already in your head? It is the easiest thing to judge from a recording, and it matters more than teams admit.
• Behavioral signal. How you handle a production incident, a disagreement, hard feedback, or an ambiguous task.
• Verbal reasoning. A few prompts ask you to talk through how you would approach a problem. They want your thought process, not a finished solution.

Read the invite carefully. If it mentions a code editor or a shared whiteboard, you are in one of the rarer technical formats and should prepare to write and narrate at the same time. If it does not, prepare for behavioral and a verbal walkthrough.

A prep flow that takes about an hour

You do not need a week. You need a focused pass.

1. Pull three or four real stories. Skim the question bank and map your own work to it. A bug that took ages to find. A technical disagreement you handled well. A time you shipped under a deadline. A piece of hard code-review feedback you acted on. Write three or four bullets per story, never a script.
2. Pick a structure and reuse it. A light STAR works for every behavioral prompt: one sentence of situation, the task, what you specifically did, the result. The STAR method on a one-way interview breaks it down line by line.
3. Rehearse two verbal walkthroughs out loud. Pick one system-design prompt and one debugging prompt and actually say them to a wall. The goal is to get comfortable narrating, not to memorize an answer.
4. Do one timed practice take. Record yourself answering one question under a real clock and watch it back once. This is the single highest-value thing you can do. Our free practice tool runs you through it with a timer.
5. Set up your space. Light from the front, camera at eye level, quiet room, phone silenced. The full checklist is in how to pass a one-way video interview.

That is it. Over-preparing past this point usually makes answers sound rehearsed, which hurts you.

The verbal code walkthrough, done well

This is the part engineers most often fumble, because the instinct from a live interview is wrong here. In a room, you can go quiet and think and the interviewer waits. On a recording, ten seconds of silence reads as a freeze.

Narrate instead. Say “let me think about the data model for a second” out loud, then think out loud. The thinking is the signal they are recording.

A reliable shape for any design or debugging prompt:

• Clarify the scope first. “Is this one region or everyone, one route or the whole site, sudden or always slow?” Stating assumptions out loud is itself a strong signal.
• Start simple, then scale. When asked to design something, reach for the simplest version that works, then say “if this needed to scale, here is what I would add.” Jumping straight to microservices, queues, and sharding on a small problem reads as over-engineering. Knowing when not to add complexity is the harder skill, and reviewers notice it.
• Name the tradeoffs. “There are a few ways to do this. Here is the one I would start with and why.” You are graded on reasoning, not on landing the single right answer.
• For debugging, measure before you fix. Walk through the layers in order and say you would not optimize anything until you had measured where the time actually goes. “Measure first” is exactly the instinct interviewers listen for.

The trap underneath all of this is answering a behavioral question like a coding problem. “Tell me about a disagreement” is asking how you handle people, not what the technically correct architecture was. Engineers often skip the human part. Do not.

About the “is it scored by AI” worry

This is the question engineers ask most, and the honest answer has two parts.

A tool can transcribe your answers, organize them, and in some setups run automated scoring to help a recruiter triage a large pool. That part is real and increasingly common. On older setups, candidates worried the camera was grading facial expressions and eye contact. One candidate wrote that HireVue “used AI to analyze your answers that you record (including things like eye contact, body language)” and worried about bias against neurodiverse and international applicants. That concern was fair.

Two things are worth knowing. HireVue has said publicly that it discontinued facial analysis in its scoring in 2021. And by recruiters’ own accounts on Reddit, most companies never switched the facial module on in the first place. As one put it, “most if not all don’t use that module.” Some jurisdictions also regulate this directly. Illinois law, for example, requires that AI-evaluated interview videos be deleted within 30 days on request.

What that means for you, practically: answer the question clearly and specifically. Do not perform for an algorithm or hold a frozen smile. A person should be the one deciding whether you advance, and clear, structured, specific answers are what survives both a human reviewer and an automated transcript. If you want the fuller picture, read whether AI can detect cheating in a video interview and do AI interviews use facial recognition.

The timer and retake mechanics

The format runs on a clock, and that is what surprises people. A typical one-way interview gives you a short prep window, usually 30 to 90 seconds, then starts recording for a fixed length, usually 60 to 180 seconds, with no pause. The prep window can be tight. One candidate described having “30 seconds to prepare for a two minute answer.” Most engineer screens run 3 to 5 questions, though some run 5 to 8.

So before you hit start, read the first screen for three things: the prep time, the recording length, and whether retakes are on. Retakes are a setting the employer controls, not a guarantee. On HireVue and similar tools the company decides whether you can re-record at all and how many times. Some allow unlimited takes until you submit, some give you exactly one with a countdown.

If retakes exist, save them for a genuinely bad take, not for chasing a perfect one. Reviewers want a real engineer, not a flawless performance, and the fifth take rarely sounds better than the second.

If you stumble

You will probably fumble at least one answer, and it matters less than it feels like it does. If retakes are off and you stumbled on a single question, keep going. One engineer posted that they had bombed a question on a recorded interview and were reassured they would still pass, which is the common pattern. Reviewers are watching whether you can think and communicate across the whole interview, not for one perfect take.

A few last setup habits that quietly cost engineers points: look at the camera lens, not your own face, because talking to the little dot reads as eye contact and watching yourself does not. Keep a few bullet points off to the side, never a paragraph to read from, because a word-for-word script reads as flat on camera. And test your microphone before the first real question, because clear audio matters more than a perfect background.

When you are ready, do a timed run on the practice tool, then walk into the real thing with the HireVue guide and the question bank fresh in your head. The format is just a different room. Once you know what it tests, the awkwardness drops fast.