</> MAANG.io

Foundations

The behavioral questions new grads and SDE I candidates are asked, from tell me about yourself to your first mistakes and disagreements. Every lesson also shows how the expected answer grows from SDE to Principal.

0 of 36 solved 0%

Tell Me About a Time You Joined a New Team and Built Trust

The question: "Tell me about a time when you joined a new team or had to work with new colleagues. How did you go about building relationships and establishing yourself as a reliable collaborator?"

Also asked as: "What did your first 90 days look like on your last team?" · "Tell me about joining a team that had no reason to trust you yet. What did you do first?" · "You've come in from outside and the org is wary of new hires. How do you earn standing?"

Who's asking: an Engineering Director or Senior Engineering Manager, in a behavioral round.

Executive Summary: The director is picturing your first 90 days on their team: what you do first, what you hold back on, how you learn what people already know, and how you can tell trust has arrived. What changes with level is what you're arriving into: a codebase, a team with its own history, an org wary of outsiders, then several orgs where your words carry further than your authority.

What the director is really asking

Underneath the story, the director is answering four questions:

  • "What will you do first?" Past behavior is the best forecast they have of your first month here.
  • "What will you hold back on?" Moving too fast and saying nothing both fail. They want judgment, not one rule.
  • "Will you find out what the team already knows?" Every odd thing has a history, and someone remembers it.
  • "How do you know it worked?" Trust shows when people bring you things they used to keep from you.

This question is about arriving. If the story could have happened in your third year on the team, pick another.

It maps most closely to Amazon's "Earn Trust"; treat that as approximate. Three ladder dimensions carry it: influence (with no credit yet), ownership (of small promises and your own early mistakes) and scope (what you arrived into).

Level fit at a glance

SDE Senior SDE Staff Principal
Asked at this level? Common Common Core Common
Story scope they expect Your first months on one team, inside one system Joining a team as its lead or a senior voice Earning standing across several teams and their managers Arriving across orgs, where whole groups act on what you say
The bar in one line Find out why things are as they are, deliver what you said you would, take correction well Learn the team's history, own a decision the team is waiting on, raise a practice Earn standing with no authority, find what crosses the teams, and get their managers to act Learn what each org and exec needs, get the decision framed in business terms, leave a mechanism
Instant down-level Coffee chats with no delivery, or a first-week rewrite told as a win "I kept my head down for months", or "I was the lead, so we did it my way" "I settled into one team and they liked me", or "my reputation came with me" "I built relationships across both orgs", with no exec decision, price or mechanism

It's Core at Staff for outside hires, who land with no credit in work that runs on influence, and Common otherwise.

Quick test: can you name one thing someone did in month three that they wouldn't have done in month one? If not, pick another story.

SDE: find out why, then deliver

What meets the bar

  • It's your first months on one team, inside one system.
  • You found out why things were the way they were, and you can name who you asked.
  • You delivered something small that you'd said you would.
  • You took correction well and changed something you do.
  • Someone did something that shows the trust: a review request, a rotation, a question brought to you first.

Model answer

"I joined the routing team at a customer-support chat company, whose software hands each customer's chat to an agent. In my third week I found a 300-line assignment function with a hard-coded 90-second wait before a dropped agent's chats moved to someone else. It looked like a leftover, so I rewrote the function, took the wait out, and called it cleanup in the pull request. The tests passed and a teammate approved it. I hadn't asked anyone why the wait was there.

It went out on a Monday, and I didn't watch the numbers. Reassigned chats jumped from about 190 a day to about 530. On Wednesday afternoon a support lead at one of our biggest customers called our account manager: her agents on poor office wi-fi kept dropping, and customers were starting over with a stranger. I reverted within the hour, about 850 extra reassignments in, and told the team at standup it was my change.

Then I asked the engineer who wrote the wait for half an hour. He'd added it two years earlier, for that customer. I added the comment and the test it had never had. I kept a list of everything that looked wrong and took it to him or our lead each Friday. Of 14 items, 11 had a reason. I told our lead I'd fix the other three by week ten, and did. One cut duplicate-notification tickets from about 30 a month to about 8.

I still thought 90 seconds was long, because a customer sits in silence when the agent has really gone. So I brought him the reconnect logs instead of a pull request: 94 percent of agents who came back did it within 45 seconds. He ran 60 seconds for two weeks, saw about 12 more reassignments a day, and kept it. In week twelve he asked me to review his next routing change, and our lead added me to on-call.

For the first month anywhere, I keep a 'looks wrong' list, and nothing on it gets a pull request until someone who was there has told me why it's like that."

Why this lands at SDE

  • "So I brought him the reconnect logs instead of a pull request": the level beat. The same opinion, held a new way, inside one system and with one colleague.
  • "Of 14 items, 11 had a reason": how they found out what the team knew, with a count.
  • "I told our lead I'd fix the other three by week ten, and did": a small promise, made out loud and kept.
  • "In week twelve he asked me to review his next routing change": trust shown by what someone did.

Follow-ups to expect

  • "Why didn't you ask before you rewrote it?": tests honesty → it looked obvious and asking felt slow. No excuse about the missing comment.
  • "How did you decide when to ask for help?": tests judgment → anything with no visible reason goes on the list; anything on fire goes to the lead that hour.
  • "What did he think of you after that week?": tests how you treat people → say what you know and how you know it. Don't guess that he was fine.
  • "What did you learn that you still use?": tests the change → watch your own release, and ask about anything with no visible reason.

Red flags

  • ❌ "I fixed their messy code in my first week and everyone was impressed." You've called the team's history a mess.
  • ❌ "Everyone was really welcoming, so it went smoothly." No test of trust, nothing delivered.

Senior SDE: own the decision the team is waiting on

What meets the bar

  • You joined a team as its lead or a senior voice.
  • You learned the team's history, including who ran things before you and what they wanted.
  • You owned a decision the team was waiting on.
  • You changed a practice for the team, or grew someone, and there's a number behind it.
  • People who disagreed had real say. Your title didn't settle it.

Model answer

"Our e-discovery product groups near-duplicates in a law firm's collected email, so lawyers don't read the same document twice. I was hired from outside as tech lead of that eight-person team. The engineer who'd run it informally for 18 months had applied for the job.

I listened first: one-on-ones, a year of incident reviews, on-call from week two. In week five the team asked me whether to patch our old grouping stage or rebuild it. She wanted the patch, because a firm may have to explain to a court how its documents were grouped. Another engineer wanted the rebuild, for the duplicates large matters left. I said I wanted to understand it better, and much the same in week seven. By week nine both were half built, and a 12-million-document matter was due in five weeks.

That drift was mine, and I told the team I'd decide that week. We replayed what each had on a closed 400,000-document matter. The patch left reviewers 8 percent fewer documents than the old stage, the rebuild 18 percent. I picked the rebuild, which shelved about seven engineer-weeks, half of them hers. She told me a replay doesn't answer a court. I agreed, kept her patch as the fallback, and asked her to write the validation standard and make the ship call against it. She held it back over attachments grouped wrongly in 0.4 percent of documents. That made the load a week late, and the firm's counsel accepted her report. Reviewers saw about 11 percent fewer against the old stage run alongside, about 1.3 million. The patch had replayed at 8, and nobody can say what it would have done live. A new matter type now waits about three days for her validation run.

No design review ends now until each open question has a named decider and a date. Over the next four months, one of nine ran past two weeks, and she decided four. On a new team, I say in week two what I'll decide by when.

Late in my first year she told me, 'For two months we couldn't tell if you had no opinion or were waiting for one of us to be wrong. Neither made anyone want to bring you anything.'"

Why this lands at Senior SDE

  • "No design review ends now until each open question has a named decider and a date": the level beat. A practice changed for one team, with a count behind it, and no other team's managers involved.
  • "asked her to write the validation standard and make the ship call against it": she gets real authority over the risk she named, and she uses it.
  • "Neither made anyone want to bring you anything": the cost of the silence, in someone else's words.

Follow-ups to expect

  • "What would she say about your first two months?": tests whether your version survives her → you've quoted her; add that half of the seven engineer-weeks were hers.
  • "How did you get the team aligned?": tests leadership → the replay, the fallback, and who still preferred the patch.
  • "Who did you grow on this?": tests people leadership → the rebuild engineer, who took a decider slot for the first time. Don't claim her: she ran the team before you came.

Red flags

  • ❌ "I was brought in as the lead, so I set the direction from day one." Authority with no history.
  • ❌ "She had trouble accepting me." Contempt, and a private matter you can't know. Say what she did.

Quick check before moving on:

Staff: earn standing, then find what crosses the teams

What meets the bar

  • You arrived with no credit, and the story spans several teams and their managers.
  • You can say what the org had already tried, and who told you.
  • You found the problem that crosses teams, with evidence people there checked.
  • Managers changed what their teams do, each for a reason of their own, with no authority from you.
  • You say "I" for the framing and alignment, credit the teams for building, and can say where the trust stops.

Model answer

"Our ad-measurement company's clients pay on audience numbers an accreditation body audits. I came in from outside to work across five pipeline teams. In week five I told three managers we should put every pipeline on one shared event platform, as at my last company. The methodology lead, there for 14 years, told me I was the third outsider to propose that, and the last attempt's re-audit had frozen method changes for a year. I'd checked none of it.

I asked my director to delay the storage review he'd hired me for by a quarter, since I had no standing yet. He agreed, and when a quarter became three he moved it to another engineer. I sat through the annual audit beside the methodology lead and read a year's 94 restated reports. I traced 63 to one cause: three teams each cut off late-arriving data at a different hour.

I showed her before any manager. She knocked out two, leaving 61. She'd known the cutoffs differed, not that they drove restatements. I held that cutoff timing wasn't a counting method. She wouldn't take my word, so she asked the auditor, who accepted it provided every cutoff was logged. With her beside me, I proposed one shared, logged cutoff to the three managers. The other two teams only read from theirs. The reporting manager signed because restatements were that team's worst weeks, the panel manager once the audit question closed. The logs manager objected that every report would go out six hours later, permanently. I showed him reruns caused 31 of his 70 night pages a quarter, and he signed. The head of sales accepted the delay, warning it could cost a client.

The teams built it over five months. Over three quarters, restatements fell from about 23 each to about 11, about $600,000 a year less in client credits. One client worth $250,000 a year left over the delay, so about $350,000 net.

Anywhere new, an old employer is never my argument, and I ask who has tried it here.

The methodology lead sends me her auditor submissions in draft now. The panel manager still copies her on everything I send him. He's watched outsiders leave and will judge me after a full re-accreditation, next spring."

Why this lands at Staff

  • "I traced 63 to one cause: three teams each cut off late-arriving data at a different hour": the level beat. A problem between teams, found by the candidate, and three managers changed their cutoffs.
  • "She knocked out two, leaving 61": the insider checks it, then asks the auditor herself.
  • "every report would go out six hours later, permanently": a cost with no end date, accepted by the head of sales.
  • "The panel manager still copies her on everything I send him": where the trust stops, with a date for the next test.

Follow-ups to expect

  • "How do you know they trusted you?": the most dangerous one → she stood beside you in that meeting and sends you her auditor submissions in draft; the panel manager doesn't yet.
  • "Why was this your problem, and what would have happened without you?": tests judgment → you were hired across the teams, and she knew the cutoffs differed but not what they drove. Credit her with the auditor and the standing. If asked whether the platform was right all along: the shared cutoff needed none of it.
  • "What did you stop doing to take this on?": tests the trade → the storage review, a quarter that became three, and the engineer who took it.
  • "How did you get the other managers to agree?": tests influence → each one's reason, and who accepted the six hours.

Red flags

  • ❌ "They were stuck in their ways, so it took time to get through to them." Contempt, and it hides the re-audit they lived through.
  • ❌ "After six months, everyone came to me." Nobody's story has everyone in it.

Principal: learn what each org needs, then get it decided

What meets the bar

  • You arrived across orgs, where groups acted on what you said.
  • You learned what each org and exec needs, from them.
  • You framed the choice in business terms, for execs, and an exec decided.
  • A mechanism other leaders must use lasted beyond the moment.
  • You can give the second-order costs in numbers, including people, and say what's still open.

Model answer

"A smart-TV maker bought a streaming-software company with about 220 engineers. I'd joined three months before the close to work across both. Two weeks after it, five of their leads asked if they'd be rewritten onto our TV software. I told them they'd keep their stack and release cycle for two years. Nobody had decided it, and it wasn't mine to give.

Four months later the CTO's integration plan merged both stacks within 18 months, saving $35 million a year, a figure he'd given the board. The leads came to me with my own words. I told them I'd had no authority, and told the CTO what I'd said. Then I asked them what those two years had been protecting. They pointed me at licensing.

They licensed their software to rival TV brands: $140 million a year, $84 million of it margin. On calls with sales, the two biggest licensees, over half of that, said they'd look elsewhere if their unreleased chip plans went through a competitor's TV teams. I put one line to the CTO and CFO: merging's $35 million wins unless it costs us more than about 40 percent of that margin, and those two alone exceed that. I drew the split: the cloud back end never sees a licensee's hardware, so it could merge; playback gets ported to each unreleased chip, so it stayed apart. The CTO took it for three years, saving $12 million, to re-decide if licensing falls below $110 million, and carried the $23 million a year gap to the board. The leads got less than I'd promised.

He'd given that number without seeing my promise or those warnings. So I proposed, and he required, one register for every decision touching the acquired org, with a named exec decider and a date, read by both sides' leads first. It ran again on our next acquisition. What I tell a new org is now labelled decided, proposed or opinion.

Licensing is at $128 million after one brand left. The two biggest renewed for one year only, so the split's on trial. The two leads who stayed send me plans before reviews. We'd planned on losing 8 of their 38 senior engineers in 18 months. We lost 14, including three of those five leads."

Why this lands at Principal

  • "merging's $35 million wins unless it costs us more than about 40 percent of that margin": the level beat. One line an exec can decide on, using the CTO's own $35 million.
  • "Then I asked them what those two years had been protecting": the people let down know what the newcomer doesn't.
  • "playback gets ported to each unreleased chip, so it stayed apart": a technical call, and it answers what the licensees feared.
  • "The two leads who stayed send me plans before reviews": the trust evidence, set beside who left.

Follow-ups to expect

  • "How did you bring the CTO along?": tests exec influence → he'd told the board $35 million, so you gave him a line he could defend and a split you designed. He decided it, and he carried the gap.
  • "What were the second-order effects?": tests whether you see past the decision → six engineers over plan, one licensee gone, the top two on one-year renewals, $23 million a year still running.
  • "What mechanism outlived your involvement?": tests lasting change → the register, why the CTO required it, and that the split is still on trial.
  • "How much of that attrition was your promise, and how much was the integration?": tests honesty → you can't separate them, and 8 was only a plan. Three of the five heard you, and all 38 lived through the merge. Say what you told those leads; don't name a share.

Red flags

  • ❌ "I reassured the team and morale stayed high." No price, no decider.
  • ❌ "Leadership broke the promise." It was your promise. Blaming execs is a no-hire at this level.

Quick check before moving on:

One story, four altitudes

SDE Senior SDE Staff Principal
Situation New on a chat-routing team, third week New lead; the informal lead had applied for the job Outside hire across five teams, third outsider with the same idea New to both sides of an acquisition
Your role Engineer learning one system Tech lead of eight Across teams, with no say over any Across orgs, with words that carried
The key move A "looks wrong" list, then logs instead of a pull request Decider and date on each design question; her ship call Traced 61 restatements to three cutoffs; three managers changed Asked the leads what the promise protected; priced and drew the split; the register
Result 850 extra reassignments, then 90 seconds to 60; review request in week twelve 11% fewer documents, a week late; 1 of 9 decisions past two weeks Restatements 23 to 11 a quarter; about $350K a year net of a lost client; six hours on every report $23M a year gap taken to the board; licensing at $128M, top two on one-year renewals; 14 left against 8 planned
Reflection No pull request until someone says why it's like that Say in week two what you'll decide by when An old employer is never the argument; ask who tried it here Label what you say: decided, proposed or opinion

Draft your answer

Step 1: Pick the right story

Test each story at your target level:

  • SDE: Did you join one team, learn why something was the way it was, and deliver something you'd said you would?
  • Senior SDE: Did you join as a lead or senior voice, keep the team's decisions moving, and change a practice or grow someone?
  • Staff: Did you arrive with no credit, did at least two other teams change what they did, and did their managers agree?
  • Principal: Did you arrive across orgs, did an exec decide something on a price you framed, and did a mechanism last?

Then check that it passes all of these:

  • You really were new, with no track record among these people.
  • It's from the last two or three years, and you were central to it.
  • Someone did something that shows the trust, and you can name it.
  • There's a number, a cost, and a mistake that was yours.

If you're a new grad. An internship, or joining a lab or open-source project midway, counts if the people there had history and code you didn't know. The director expects a codebase and a few months, but still wants what you learned before changing anything, something you delivered as promised, and one thing someone did that showed trust: "The lab's simulation code was four years old, and the PhD student who wrote it answered questions with 'read the code'."

Aiming a level above your current title? A Senior aiming at Staff can only tell the cross-team version if other teams' managers changed something after you arrived. Becoming your own team's go-to person isn't that. If no story clears the level, tell your best story at your real altitude, and use the reflection to show you see the next one: "If I joined your org, I'd spend my first month on what crosses the teams, and I'd ask who has tried it already."

Step 2: Fill the skeleton

SITUATION        (15s) — What you walked into, who was there first, and what
                         they'd seen before: ______
FIRST MOVES      (15s) — What you did first, and what you held back on: ______
WHAT THEY KNEW   (10s) — How you found out why things were that way, and from
                         whom: ______
ACTIONS          (30s) — 2–3 moves, each "I …, because …". Who moved, and why: ______
LEVEL BEAT (inside ACTIONS) — the line only your level could say.
                         SDE: what you asked or checked before acting on your
                         own view. Senior: the practice you changed, or the
                         person you grew. Staff: what several managers changed,
                         with no authority from you. Principal: the price execs
                         decided on, and the mechanism others must use.
YOUR MISTAKE     (15s) — What you got wrong as the new person, with no "but": ______
RESULT           (15s) — The number, net of what it cost: ______
TRUST EVIDENCE   (10s) — What someone did that they wouldn't have done in month
                         one and, if true, who still doesn't trust you: ______
REFLECTION       (10s) — What you do differently when you arrive somewhere: ______

Aim for about two minutes, two and a half at most, with a 30-second cut ready (what you walked into, what you did first, one number, how you know).

Step 3: Calibrate it

If your answer says… It reads as… Fix
"I listened for my first 90 days" Passive, with no plan The date you said you'd have a view
"At my last company we…", as your argument An outsider who hasn't looked What you learned here, from whom
A Staff story inside one team Senior The other teams, and what their managers changed
"I built relationships across both orgs", told at Principal Staff or below The exec decision, the price, the mechanism

Step 4: Rehearse the follow-ups

Prepare these five, and say each answer out loud once:

  1. "How do you know they trusted you?" The most dangerous one. Name what someone did that they wouldn't have done earlier, and who still hasn't.
  2. "If you joined my team on Monday, what would you do in your first 30 days, and what wouldn't you touch?" Who you'd ask, what you'd deliver, and one thing you'd leave alone.
  3. "What did you hold back on, and what did that cost?" Holding back has a price too. Give it.
  4. "Who was there before you, and what did they want?" Their reason, with no pity and no scorn.
  5. "What would the person least happy about your arrival say?" Their words if you have them, and what you did about it.

Common mistakes

  • Using the old team's flaws as your setup. Give their reasons.
  • Ending at month three, above SDE. Say what happened later, and who hasn't come round.

Key takeaways

  • It's a forecast of your first 90 days: what you do first, what you hold back on, and who you ask.
  • The level shows in what you arrived into: a system, a team, an org, then two orgs and an exec decision.

Test yourself on this question's key ideas:

Sign in to MAANG.io

Use your Google or Microsoft account — no password to remember.

Continue with Google Continue with Microsoft

Please accept the terms above to continue.