← All topics
🗣️ Talking About Your Own Work

Interview Execution Tips

How to handle the moments that aren't about knowing the answer.

When asked to refactor code live (like the DiscountService exercise): narrate your thinking as you go — name the smell you see first ("this if/else means every new customer type needs a code change, that's an Open/Closed violation"), THEN write the fix. Interviewers are evaluating the reasoning process at least as much as the final code.

When you don't know something: say so directly, then reason toward the likely answer from what you do know, out loud. "I haven't used NServiceBus sagas specifically, but based on how sagas work in other frameworks I'd expect X" is a strong answer. Bluffing and getting caught is far worse than an honest gap paired with good reasoning.

When a question is ambiguous: ask a clarifying question before answering — "when you say handle errors, do you mean transient failures or a message that will never succeed?" mirrors exactly the distinction in the Retry/DLQ topic and signals you're already thinking at the right level.

Connect answers back to their stated stack when relevant: if asked generally about testability, land on "...which is exactly why the client's preference for DI/composition over inheritance matters — it's what makes this kind of code mockable in a unit test." Showing you've internalized their stated architectural preferences, not just generic best practice, is a strong signal you were paying attention to the brief they sent you.

Flashcards (2)

When live-refactoring code in an interview, what should you say BEFORE writing any new code?
tap to reveal answer
Name the specific smell/problem you see in the current code and which principle it violates (e.g. 'this if/else chain violates Open/Closed because every new case needs a code change') — interviewers weight the reasoning process, not just the final diff.
What's a better response than guessing when you genuinely don't know something in a technical interview?
tap to reveal answer
Say directly that you haven't worked with that specific thing, then reason toward a likely answer out loud from adjacent knowledge you do have — demonstrates honesty and problem-solving over a bluff that risks being caught by a follow-up question.