A senior engineer strongly opposes your architectural decision. You have decided anyway. How do you handle it?

What the interviewer is testing

Whether you can close a disagreement in a way that produces committed implementation rather than passive resistance.

Before deciding

Steelman their position and check you understand it. This is not politeness — it frequently reveals that the disagreement is about an unstated constraint one party can see and the other cannot, and the real resolution is there.

Separate the disagreement into facts, forecasts and values, because they need different treatment. A factual dispute is testable. A forecast can be made concrete by asking what would confirm or refute it. A values disagreement is a genuine difference in priority that must be surfaced and escalated rather than argued.

Make it testable if you can. A spike, a load test or a prototype settles more arguments in two days than a fortnight of debate, and both parties can accept the result without loss of face. If the disagreement is resolvable by evidence, get the evidence — it is almost always cheaper than the argument.

After deciding

Explain the reasoning, including what you weighed and why their concern did not outweigh it. A decision explained is easier to commit to than one announced. Do not pretend the concern was invalid.

Record their reservation in the ADR, named, with the reasoning. This is what makes disagree and commit honourable rather than suppressive: it is a documented position that can be revisited with evidence, not a grudge.

Agree what would trigger a revisit — a specific signal, a metric, a date. That gives the disagreement a legitimate future rather than requiring them to be permanently wrong.

Then ask for genuine commitment, and expect it. Half-hearted implementation of an architecture is worse than either option done properly.

If they are more likely right than you

Say so and change the decision. Being wrong publicly costs far less than a bad architecture, and it is the single most effective thing an architect can do for their credibility.

What a strong answer adds

Knowing who decides. An argument with no decision-maker continues until someone tires, which selects for persistence rather than for the better answer — so establishing the decision right explicitly is part of the resolution.

Common weak answers

Overruling on authority. Deferring indefinitely to avoid conflict.