Where the review round shows up.
The landing page says code review rounds are run at Airbnb, Meta and Google. This page is the evidence for that claim — what each source actually says, when it said it, and how much weight we put on it.
Read this first
Several large companies, Airbnb among them, publish little or nothing official about their interview loops. Every claim below rests on publicly posted candidate reports and third-party interview guides, not on company statements. Interview formats change; a report from 2024 may not describe a loop in 2027.
None of these companies is affiliated with, sponsors, or endorses Code Review Wars. Their names and logos appear only to identify where the interview format exists.
All sources accessed 13 September 2026 · Page last reviewed 13 September 2026
Airbnb
Claim
Airbnb runs a dedicated, rubric-scored code review round in its software engineering loop.
Evidence
- Blind · Reply dated 5 April 2026“Interviewing at Airbnb” — community thread, reply from a candidate ↗
“They have something unique called code review rounds where an engineer silently watches you review code while sharing screen. They usually have 3 pull requests to review here and first one is too small in scope and very easy.”
- Blind · 27 August 2024“AirBnb Senior Onsite Interview Rejection” — candidate write-up ↗
“Code Review: You get a GitHub repo with 3 pull requests. What they don’t tell you is that you score ‘points’ for making comments on various issues about the code.”
- Exponent · Updated mid-2026“How to Ace a Code Review Interview” — guide quoting an Airbnb interviewer who helped design the round ↗
Describes a formal, rubric-scored code review round used from senior through staff level: three questions drawn from an approved bank in Java or Python, scored on communication, teamwork, execution and depth of knowledge rather than raw issue count.
- Hello Interview · 2026 editionAirbnb G9 (senior) interview guide ↗
“Airbnb introduced this round in 2024 to replace the traditional second coding interview.” Describes a 60-minute round reviewing pre-written code for bugs, edge cases and maintainability.
Strength
Strong. Two first-hand candidate accounts plus two independent interview guides, one of which quotes an Airbnb interviewer who helped build the round.
Meta
Claim
Meta has used code review as an interview format.
Evidence
- Exponent · Updated mid-2026“How to Ace a Code Review Interview” — compiled candidate accounts ↗
A Meta production-engineering candidate reports that code review rounds substitute for coding rounds in their loop, and that Meta now allows an AI code assistant in the round.
- Hello Interview · Rollout began October 2025“Meta’s AI-Enabled Coding Interview: How to Prepare” ↗
“In October 2025, Meta started rolling out a new interview type called AI-enabled coding.” Candidates work in “a multi-file project, with existing classes, data models, and logic already written. You didn’t write any of this code.”
Strength
Strong for production-engineering loops, where a candidate reports code review substituting for coding rounds. We do not claim it is universal across every Meta level or track. Meta’s newer AI-enabled round tests reading an existing codebase but is not a pull-request review.
Claim
Google’s loop includes a code comprehension / debugging exercise that tests the same skill: reading someone else’s code, finding what is wrong, and fixing it.
Evidence
- Exponent · Updated mid-2026“How to Ace a Code Review Interview” — compiled candidate accounts ↗
An L6 engineering-manager candidate was given “more than 200 lines of broken code in a Google Doc, with no IDE and no AI tools” and asked to review, comment, fix defects and reason through test cases by hand.
- Exponent · Updated 2026“Google’s AI-Assisted Coding Interview (2026 Guide)” ↗
Describes a code comprehension round built on “reading, debugging, and optimizing real code with Gemini available as an AI assistant”, piloted for junior and mid-level roles on select US teams.
Strength
Moderate. Google’s format is a debugging exercise on broken code — not a pull-request review. The skill under test is the same; the format is not identical, and we say so.
Apple — assessed, not shown
In the same Exponent compilation, an Apple M1 engineering-manager candidate's round labelled “code review” was a from-scratch coding problem on CoderPad (validating a completed Sudoku grid) followed by an optimisation discussion. One account of a coding problem is not evidence of a code-review round, so Apple is not on the landing page.
Spotted an error, or have a first-hand account we should add? support@codereviewwars.dev