Modesty reads as a weak answer, and it should not
Mandarin has no tense, and the interview runs on tense
Mandarin marks time with words — 昨天, 已经, 了 — and leaves the verb alone. English marks it on the verb, obligatorily, every single time, and then builds its most important interview question on that marking. “Tell me about a time when” is a request for the past tense specifically, and it is heard as wrong the moment the verb is bare.
Modesty that is read as a weak answer
谦虚 is not shyness and it is not false modesty. Downplaying your own part is how a competent person signals respect for everyone else’s, and in most Chinese workplaces the person who does it is trusted more, not less. It is a genuinely better social norm than the one an American interview runs on.
The problem is that a behavioural interview is scored, usually against a rubric, by someone who has 40 minutes and no other source of information about you. “I just did a small part” is taken literally, because there is nothing else to take it as. The interviewer does not think you are being modest; they write down that the candidate’s contribution was small.
The softeners that stack
“I think maybe I could perhaps have helped a little with the design.” Each hedge is polite on its own; four in one sentence removes the claim entirely. One hedge is plenty in English.
Stating a fact is not boasting
“I designed the caching layer; it cut p95 latency by 40%” contains no self-praise. It is a description with a number in it, and it is what the question was asking for.
“We” when they asked about you
我们 is the natural subject for work that a team did, and in Chinese saying 我 repeatedly about a shared achievement sounds arrogant. Carried into English, every sentence starts “we”, and the interviewer — who asked “what did you do?” — has no way to separate you from the four other people in the story.
The fix is not to become boastful. It is to split the sentence: “The team migrated the whole platform. I owned the data layer and wrote the rollback plan.” The team keeps its credit, and your verb is visible. Two other patterns sit next to this one. The first is answering “tell me about a time when” with a policy — “normally I would first check the logs” — which is a good answer to a question that was not asked; they want one specific day. The second is 面子: “tell me about a failure” is genuinely uncomfortable, and an answer that avoids the failure reads as evasion. The expected shape is a real failure, stated plainly, followed by what changed because of it.
The phrases that ask for a story
“How did you land on that?”
Nothing to do with landing. It means: how did you arrive at that decision, what were the alternatives, why this one. The answer they want has a rejected option in it.
“Tell me about a time when…”
A request for one specific past event, with a date, people and an outcome. The most common misfire is a general description of how you usually work, which scores as no answer at all.
“What would you do differently?”
Not an accusation and not a trap. “Nothing” is the weakest available answer. They are testing whether you can assess your own work, which is a separate skill from doing it.
“Sell me on it”, “what’s your gut feel?”
The first asks you to argue for a position out loud. The second asks for an opinion without evidence, on purpose — and “I would need more data” is a real answer only if you then give the instinct anyway.
One structural habit is worth keeping rather than losing: 第一、第二、第三. Numbering your points is clear, and English-speaking interviewers like it. Just make sure the numbered list is inside a story, not instead of one.
What the window gives you
Cue is a Chrome extension. You click its icon in the tab the interview is in, and Chrome gives that tab’s audio to it — nothing joins the call, no participant appears, and the microphone is never requested. It runs in Google Meet, Zoom on the web and Teams on the web only.
The interviewer’s line appears as it was recognised, and under it the same line in Chinese if that is what you picked from the 68 languages in the picker. Translation is on from the first minute and lives in two buttons in the window header; set both to the same language and it stops. When the copilot judges that a question was asked of you, a card appears — short, structured, written in the language you chose to read, with the line it grew from attached so you can confirm it is answering the right thing.
Two features matter specifically for what this page is about. The note you write before the call — role, team, the two projects you want to reach — is what makes a card say “the caching layer you built” rather than a generic paragraph. And the box for your own question means you can type “what does land on mean here” mid-round and read the answer without stopping the conversation.
Where it stops
It cannot correct your English while you talk, because it does not hear you. It makes no recording, no summary and no notes, and there is no way to give it a CV or any other document. Recognised text and answers are kept 30 days and then deleted; audio is never stored at all.
And it cannot supply the thing this page is really about. Saying “I designed it” out loud, in a room where that feels immodest, is a decision you make. The card can put the sentence in front of you. Speaking it is yours, and so is the follow-up question that comes after it.
FAQ
This language, specifically.
Can I read the interviewer in Chinese and get the card in Chinese too?+
Yes, and that is a common setup. The call is recognised in English, the line under it is shown in Chinese, and the card is written in Chinese so you read it at full speed. Some people prefer the card in English instead, so the phrasing they are about to say is already in front of them. Both are switchable mid-call.
Does the tense problem affect what it understands?+
No — it is listening to the interviewer, not to you. Your own grammar never reaches it, because the microphone is never requested. What reaches us is a window of the last twelve recognised lines from the call itself.
I answer “tell me about a time when” with a general description. Can the card catch that?+
Partly. The card is built around a specific answer shape, so it tends to come back asking for the event rather than the policy — and the more you put in the pre-call note about actual projects, the more specific it is. Noticing the pattern, which is what this page is for, does more than any feature.
Will it help with a fast American interviewer on a poor connection?+
That is one of the two things it genuinely does well. The recognised line stays on screen after the sentence has passed, and a strip of recognised speech runs under the window, so a swallowed word is visible rather than simply lost.
Does it work for interviews conducted in Chinese?+
Chinese is in the picker on both sides, so a call in Chinese is recognised and the card can be written in English or Chinese. But the product and the prompt behind it are built for interviews, in whichever language — not for a translation tool covering an entire meeting.
Let the round be about your work.
Trippi Cue is in the Chrome Web Store. The interviewer arrives in your language, and your answer is written in the one you pick.