Is Their Japanese Enough? How Employers Judge Engineers

The JLPT N1 tells an employer that an engineer can read and listen to difficult Japanese. It tells them nothing about whether that engineer can push back in a design review. The test has no speaking or writing section, and the organizers say they have no plans to add one. So when hiring managers in Japan ask whether someone's Japanese is enough, the certificate answers only part of the question.

What does N1 actually prove?

That the person can understand demanding Japanese, on paper and by ear. The official description of N1 is the ability to understand Japanese used in a broad range of situations: newspaper editorials and critical essays on the reading side, natural-speed conversation, news and lectures on the listening side.

Passing takes two things. The total score has to reach the pass mark, 100 for N1. And each of the three sections, language knowledge, reading and listening, has to clear its own minimum, 19 points each for N1. Miss one section minimum and you fail, however high the total. So an N1 does rule out a glaring weak spot in those three areas.

It does not test speaking or writing at all.

The test's organizers, the Japan Foundation and JEES, also publish a "Can-do" list (2020) built from a 2012 self-assessment survey of 64,842 people. A statement goes into a level if at least 75% of people who passed that level said they could do it, or just about. One N1 statement reads, roughly: "Can take part in a discussion on a topic of interest and state an opinion logically." Two phrases matter there. It is self-assessed. And it is a topic of interest. A meeting where a product manager wants to cut your design is not always one.

The situations employers actually look at

Employers care about the specific jobs the engineer will do in Japanese once they join. The ones that come up most:

• Reading specs, tickets and customer request emails, and knowing what to ask afterwards.
• Writing: incident reports, pull request descriptions, weekly updates. Japanese companies expect more from written work than many newcomers assume.
• Disagreeing in a meeting, on the spot. Hearing that "this might be a little difficult" means no, and being able to say "this design carries this risk after launch" without making anyone lose face.
• Explaining to non-technical colleagues why something was built the way it was.

Which ones matter depends on the team. In a team that works in English, Japanese may only come out at lunch. In an in-house product team with Japanese product managers, the last three happen daily. In a role placed on site at a client, all four matter, and the audience is the customer.

N1 covers the left side, reading and listening; employers worry most about the bottom right, pushing back live in a meeting

Four situations where engineers use Japanese at work, split by input or output and by prepared or live. The JLPT tests language knowledge, reading and listening, and not speaking or writing (JLPT official FAQ, as of November 2026). The grouping of situations is this article's own

Across: input (read, listen) to output (write, speak)　Down: prepared to live

Reading specs and tickets

Input, prepared. Largely covered by N1 reading

Writing incident reports and PR notes

Output, prepared. Not tested by N1

Hearing what a meeting really means

Input, live. Partly covered by N1 listening

Pushing back in a meeting, live

Output, live. Not tested by N1; the one to test in interview

The certificate can vouch for the left two. The right two have to be seen in the interview.

How to test it in an interview

Have the candidate do the work once, in the room. Careful employers set it up something like this.

Reading. Hand over a redacted page of a real spec, give ten minutes, then ask what they would want to check with the product manager. The quality of the questions tells you more than reading speed.

Writing. Give a made-up outage and a timeline, and fifteen minutes to write a short incident report. Look at structure, whether the honorifics hold together, and whether someone who wasn't there could follow it.

Speaking. The interviewer proposes a design with a deliberate flaw and watches whether the candidate can point it out in Japanese, give a reason, and keep the room friendly. This is the hardest one to fake.

Score each situation separately. Don't collapse it into "Japanese: good." This follows the logic of Does this person fit? Veto, weakest layer, five levels. Language sits in the equipment layer. First separate the situations the role needs on day one from the ones that can be built after joining. Then let the weakest of the day-one situations set the result, with no averaging. Anything the interview didn't test gets marked as unverified, and the N1 doesn't fill that gap.

It works the other way too. A candidate without N1 who can argue a design in Japanese at a whiteboard should not be screened out for lack of a certificate. I made a related point about multilingual staff in Your bilingual staff are not translators: look at what someone gets done with a language.

What evidence helps an employer decide

A record of the situations. "N1" on a CV says the person cleared a bar. "Ran the weekly client meeting in Japanese" or "owned incident reports for the team" says they did the job. Doing it live in the interview is better still.

N1 still has value. It saves a screening round, and Japanese ability earns points under Japan's points system for highly skilled professionals. It just answers a slightly different question from the one the hiring manager is asking.

I am not sure how long the reading test will matter. Translation tools are getting good enough that reading specs may stop separating candidates in a few years. I can't think of a tool that will push back in a meeting for you.

---