Mercari interview, decoded.

Three to six weeks is the usual span and candidates report about three weeks on average, longer when the take-home sits in the middle. The stages below come from Mercari's own recruitment page, with the assessment detail from candidate accounts.

Two things make this loop preparable in a way most are not. The assessment is consistent enough that candidates describe the same assignment, and the interview method is published by Mercari itself, down to the probes under each letter of STAR and the values they map to. Treat the first as an engineering exercise with a known shape and the second as an exam with the syllabus printed. What follows separates what Mercari says from what candidates report, because only one of those is official.

• The assignment candidates describe: About eight hours of work: a pub/sub web API built to scale to roughly 200,000 requests a second, with high unit-test coverage and documentation. Mercari's own page says only that there is a technical assessment for certain positions, so treat the specifics as the reports they are and confirm the brief you are given.
• How to spend the eight hours: Two on the design and the interfaces, four on a working core with tests, one on the documentation, one on the load argument. A reviewer reads the README first: state what you built, what you deliberately left out, the delivery guarantee you chose and why, and the number you think it holds. A smaller system that is correct and explained beats a larger one that is neither.
• What the reviewers press on: Delivery semantics (at least once against at most once, and what a duplicate costs), back-pressure when consumers fall behind, what happens on restart, and where the first bottleneck sits at the stated load. Have a number for each, even a rough one, and say how you would measure it.
• STAR, as Mercari prints it: Situation: the project, its goal, what it looked like at the start, who was in the team and how hard it was. Task: your role, why it was yours, and the challenges. Action: what you did, how many times, over how long, to whom, why that and what else you considered. Result: what came of it. Prepare two projects to this depth; the follow-ups are designed to go past the first layer.
• The values inside the questions: Mercari says the Action probes are also value probes: how bold the action was (Go Bold), how much you involved others (All for One), the expertise you showed (Be a Pro), and how fast you moved (Move Fast). Pick stories where one of the four is unmistakable, and know which is which before the round.
• The language question, answered honestly: Applications are accepted in English or Japanese, English is the working language across much of engineering, and some teams require a specific language, which the posting will say. Candidate accounts of Japanese-language interviews exist for some roles, so ask the recruiter which language your loop runs in rather than assuming either.
• Moving to Japan: The relocation package is published: visa sponsorship and support, a one-way flight, shipping, the first month's accommodation, a moving allowance, and help with residence registration, a bank account and a phone. Support is stated for partners and dependants, including same-sex and de facto partnerships, which Mercari asks you to raise early in the process.

Build the take-home before it arrives: give yourself one evening and a pub/sub API with tests, a README and a load argument, then throw it away and do it again with what you learned. Then write your two STAR stories out in full against Mercari's own probe list, and name the value each one demonstrates. That is most of this loop.

Mercari publishes four values and says outright that its interviewers ask about them inside the STAR questions, so each one is worth a story before you sit down.

• Go Bold: the interview asks how bold the action you took actually was, so pick the decision that could have failed
• All for One: how much you involved other people and other teams, rather than how much you carried alone
• Be a Pro: the expertise you brought as a professional in your field, demonstrated in the detail
• Move Fast: what you did quickly, and what you were willing to leave imperfect to do it
• Enthusiasm for the mission, which Mercari names as one of the three things every interview is judging

Mercari prints the shape of its own interview, which is unusual enough to prepare against directly. The questions below are the probes its recruitment page names, plus what candidates report from the technical rounds.

Levels.fyi for Mercari software engineers in the Greater Tokyo Area, read on 23 September 2026, from a modest sample; the same pages show a lower median for all reports, so treat these as a band. Japanese packages are mostly cash, and the numbers are well below a Bay Area offer for the same level: the arithmetic that matters is local, since Tokyo rent, tax and the cost of living differ as much as the salary does. Mercari is listed in Tokyo, so equity in an offer is priced publicly. Ask what the grant is, whether the level is set before or after the loop, and what the relocation package covers in your own case.

1. Build a pub/sub API with tests, a README and a load argument once before the take-home arrives, and time yourself.
2. Write two projects out against Mercari's published STAR probes, then label which of the four values each one demonstrates.
3. Have an answer to "what are our values", because candidates report being asked it directly.
4. Ask the recruiter which language your loop runs in, and whether your position carries the technical assessment.
5. Prepare the why Japan answer as seriously as the why Mercari one: both come up, and the first is the one candidates fumble.
6. Read the relocation package before the offer conversation so you can ask about the parts that apply to you.

This guide covers Mercari's engineering and research hiring. For management and leadership roles the loop is similar but the bar shifts to people, delivery and strategy, so pair it with the leadership interview prep hub. The bar for your exact role comes from the role-by-role guides, and the prep that actually transfers is spoken, so run a mock interview before the real one.

Also interviewing at Rakuten, Coupang or Samsung R&D India? Their guides follow the same format, from the first screen to the offer.

• 01Mercari Careers, Recruitment and Selectionthe primary source: the selection process (document screening, technical assessment for certain positions, interviews, offer meeting, reference check), the three things interviews judge, the STAR method and the probes under each letter, the four values inside the Action questions, and the English or Japanese application policy.
• 02Mercari Careers, Work in Japanthe relocation package: visa sponsorship, flight, shipping, first month's accommodation, moving allowance, settling-in support, and the stated support for partners and dependants.
• 03Glassdoor, Mercari software engineer interviewcandidate accounts of the take-home: about eight hours, a pub/sub web API to roughly 200k requests a second with unit tests and documentation, and the rounds that follow it.
• 04Levels.fyi, Mercari software engineer, Greater Tokyo Areathe MG1 to MG4 figures and the Tokyo median, read 23 September 2026.

Interview processes change. This reflects widely-reported and sourced conditions as of 2026 — confirm specifics with your recruiter, and treat it as a map rather than a guarantee.