Feedback and Difficult Working Relationships

We had different communication styles, so I stayed professional and found ways to work around them.

This sounds restrained. It is often a story about avoidance. The interviewer cannot tell what the other person did, whether it harmed the work, what you said, or why the relationship would be any better on the next project.

A difficult-relationship answer should not ask the interviewer to decide who was the difficult person. It should show whether you can be trusted with tension. Can you replace a character judgment with evidence, say the hard thing without humiliating a colleague, hear inconvenient criticism of your own work, and leave behind a working arrangement that survives the conversation?

Find the next encounter

Candidates often end these stories with “we talked it through.” That is where the useful part begins. What happened in the next design review, on-call handoff, planning meeting, or code change? Did either person behave differently? Did the work improve? Did the agreement hold when pressure returned?

Choose a story in which you had to preserve both a standard and the ability to work together. There should be enough repetition to distinguish a pattern from one bad afternoon, but not months of silent resentment. Protect identifying and sensitive details. The interviewer needs the conduct and its consequence, not a colleague’s name, health information, or performance history.

Avoidance and overreach are the two easy failures. Avoidance leaves the standard implicit while you route work around the person. Overreach turns a peer into an amateur manager who diagnoses motives, promises a support plan they cannot own, or treats escalation as punishment. The useful territory lies between them: act directly within your authority, and recognize the boundary of that authority.

The reviews everyone learned to endure

Consider a modeled example. A senior engineer is incisive in design reviews and often finds the risk that others have missed. They also interrupt presenters before a trade-off is complete and dismiss questions with “we already know that will not work.” Less experienced reviewers have started sending their concerns privately after the meeting. In one case, a backward-compatibility problem surfaces only after implementation begins.

The candidate chairing these reviews initially compensates. They ask presenters to rehearse, collect questions in direct messages, and schedule follow-ups for decisions that the room could have made. The meetings become quieter, but the workaround protects the behavior and makes review slower. Calling the peer “abrasive” would be satisfying and nearly useless. It says nothing the peer can verify or change.

Before speaking to them, the candidate reconstructs two reviews:

• the presenter was interrupted before explaining the migration boundary;
• two reviewers stopped trying to ask their questions after being cut off;
• the compatibility concern appeared the following day, after the design had been approved; and
• the candidate’s private collection of questions concealed the loss of signal from the room.

These observations do not prove a motive. They do establish a work problem. The distinction protects both people: the feedback can be direct because it does not depend on a verdict about personality.

Let the conversation carry evidence in both directions

The candidate speaks privately and stays close to the work:

In the last two reviews, you interrupted the presenter before they finished
the migration trade-off. After that, two reviewers sent concerns to me rather
than raising them in the room, and we missed a compatibility issue until the
next day. I want the review to keep the rigor you bring without teaching
people that questions are unwelcome. What are you seeing that I may be
missing?

The final question is not decorative softness. The peer has relevant evidence: recent proposals have arrived without migration plans, and the candidate has allowed meetings to become live discovery sessions. The peer says, sharply, “You are asking the room to rescue half-finished documents.”

Now the candidate is receiving difficult feedback inside a story about giving it. Defending their tone would be an escape. Accepting the accusation whole would be equally careless. The candidate asks for two examples and what a review-ready proposal would have contained. One example supports the criticism; the other was complete but unconventional. The candidate owns the real part: unclear entry criteria have made reviewers do expensive work in the meeting. They do not concede that interruption is therefore harmless.

This is often the most credible turn in a feedback story. The other person is not a prop, and the candidate does not become correct merely by initiating the conversation. Each person can contribute to the strain without their contributions being identical.

Build an agreement the work can test

“We will both communicate better” cannot survive contact with the next deadline. The pair instead change the review itself.

Authors will circulate a decision-ready document forty-eight hours before the meeting. It must state the compatibility boundary, migration path, and open decisions. The candidate will return documents that are not ready rather than asking the room to repair them live. Reviewers will put blocking concerns in the document with evidence. During the meeting, the presenter gets to finish the trade-off before objections begin, unless continuing would create an immediate safety or incident risk.

The agreement has reciprocal obligations, but it does not manufacture equal blame. Better documents address the valid criticism of the candidate’s process. Letting people finish and treating questions seriously address the peer’s conduct. The work now has a way to reveal whether either change helped.

At the next review, the proposal arrives with the migration boundary exposed. The peer writes a blocking concern in advance: the compatibility plan assumes all consumers upgrade within one release window. The author revises the plan before the meeting. During discussion, a newer engineer asks whether old and new payloads can coexist. The peer waits, then extends the question with a failure case rather than dismissing it. The review is still demanding. It is also shorter, and the decision record contains the concern that would otherwise have surfaced during implementation.

That is a stronger outcome than “we became friends.” A working relationship need not become warm to become candid, predictable, and useful.

When the agreement bends under pressure

One successful review is encouraging, not conclusive. Several weeks later, an urgent proposal arrives late. The peer again talks over the presenter. The candidate pauses the meeting and restates the review rule without conducting a public trial: the presenter will finish; blockers will then be tested. After the meeting, the candidate and peer distinguish two failures. The proposal should not have entered review in that state, and the deadline did not cancel the agreement about how people participate.

Because document readiness affects the whole team, the candidate asks the manager to sponsor explicit entry criteria and a facilitator for high-risk reviews. This is not a disguised request to punish the peer. It moves a team mechanism to the person who can establish and enforce it. The candidate tells the peer what they are raising and why, then follows up afterward to confirm how the next review will run.

Escalation becomes credible when you can name the threshold that peer influence crossed. Perhaps clear feedback did not change a repeated pattern. Perhaps role expectations, staffing, or a support plan are now at issue. Perhaps the effect reaches people whom a private agreement cannot protect. State what you tried, what continued, what work was at risk, and what decision the manager or formal owner needed to make.

Do not confuse this threshold with severity that requires an earlier route. Harassment, discrimination, threats, retaliation, security concerns, and immediate safety risks may belong with a manager or the organization’s formal reporting, security, or incident process from the start. Follow the applicable policy; do not attempt solo mediation or promise confidentiality you may be unable to keep. An interview answer should show that you know the limit of ordinary peer feedback, not that you can recite employment policy.

Support a colleague without becoming their hidden manager

Questions about a struggling colleague test the same ownership judgment. A peer can share context, pair on a bounded task, provide an example, review a plan, or give direct feedback about a missed handoff. A peer usually cannot set formal performance expectations, decide accommodation, control staffing, or commit the team to indefinite rescue.

A useful offer names its edge:

I can pair on the first migration and review the next two plans. I cannot own
each migration behind the scenes, because then planning never sees the real
capacity or support need. Let us tell your manager what we are trying and use
the next two plans as the checkpoint.

The boundary is not a threat. It makes the support honest to the colleague, the team, and the manager who owns the broader expectation. In a story, show what help you gave, what the colleague still owned, what evidence you expected to change, and when you involved the appropriate owner. Do not turn private performance details into proof of your leadership.

Repair is part of the result

After a forceful conversation or escalation, silence leaves both people to invent what the relationship now means. Repair makes the next collaboration less ambiguous. It may be as simple as acknowledging a poor assumption, checking that the agreement was heard the same way, or giving credit when the other person changes their practice.

In the review example, the candidate tells the peer that their criticism of document readiness changed the intake process. They also say plainly that raising the team norm with the manager was necessary after the interruption recurred. This is neither apology theater nor a retreat from the standard. It allows candor to run in both directions without making future work depend on unspoken resentment.

Be precise about the outcome. “Trust was restored” is usually too large to prove. “They raised blocking concerns in writing, newer reviewers participated again, and we could challenge each other’s proposals without a second meeting” is evidence an interviewer can inspect.

Let the interviewer test your judgment

Stay inside one event when the interviewer presses. The pressure often reveals whether your care, candor, and boundaries were real.

Interviewer: Why did you not go straight to their manager?

Candidate: I had two concrete examples of peer behavior that I could address
directly, and there was no safety or conduct issue requiring a formal route.
The first conversation also exposed a problem I owned: weak review entry
criteria. I wanted a work agreement we could test before asking a manager to
own it.

Interviewer: Did you tell them they were intimidating junior engineers?

Candidate: I did not claim to know how everyone felt. I said that two
reviewers stopped asking questions after being interrupted and sent the
questions privately, which cost us review signal. That was observable and
consequential.

Interviewer: What criticism did you receive?

Candidate: They said I was asking meetings to rescue incomplete proposals. I
asked for examples. One showed that our intake standard was genuinely unclear,
so I changed it. Another proposal was complete, so I did not accept that as a
reason to cut off its author. The conversation changed my process without
erasing the feedback I needed to give.

Interviewer: Why escalate later?

Candidate: Under deadline pressure, the interruption pattern returned, and
document readiness had become a team-wide inconsistency. Peer agreement was
no longer enough. I asked the manager to own the review entry criteria and
facilitation rule; I did not ask them to adjudicate who was nicer.

Interviewer: How did you know the relationship improved?

Candidate: The next reviews gave us observable evidence: blockers appeared in
the document earlier, presenters completed their reasoning, and newer
reviewers raised concerns in the room. We still disagreed, but we no longer
needed workarounds to use each other's judgment.

The answer contains no personality diagnosis and no claim of perfect harmony. It shows a pattern, a conversation that corrected both sides, a test in real work, a recurrence under pressure, an ownership boundary, and a usable relationship afterward.

Rebuild one feedback story

Choose a story from your bank and reconstruct the sequence before rehearsing it:

1. Replace every character label with behavior you observed and a consequence for quality, delivery, customers, operations, or participation.
2. Identify the part you owned, including valid criticism you were initially reluctant to hear.
3. Write the first two sentences of the private conversation. Include the behavior, the effect, and a genuine question about missing context.
4. Define what each person agreed to do differently in the next real interaction.
5. Name the evidence that showed improvement or recurrence.
6. State the boundary that would move the matter to a manager, formal channel, security process, or incident owner.
7. Show the repair: what made later collaboration possible without pretending the tension vanished.

Rehearse the answer in about two minutes, then invite interruption. Useful probes are: What did you get wrong? Why was direct feedback appropriate? What changed in the work? Why did you escalate when you did? What help was yours to offer? What detail are you withholding to protect the other person?

The aim is not to perform perfect composure. It is to show disciplined care when the work and the relationship are both under strain. Senior judgment is visible when candor produces a better next encounter—and when you know that some encounters should not be yours to manage alone.

Related links

• Conflict and Technical Disagreement
• Mentoring and Raising the Engineering Bar
• Ownership and Ambiguity
• The Senior Story Spine
• Confidentiality and Sanitization