Behavioral: Resolving a Workplace Conflict and Using Generative AI in Your Work

In the behavioral part of a software engineering onsite, you are asked two questions: one about a conflict you worked through, and one about how you have used generative AI. Answer each with one specific story from your own experience. ### Clarifying Questions - Does a disagreement with a partner team or a manager count as a conflict, or should it be with a peer on your own team? - Does the generative AI question mean using AI tools to do your engineering work, or building product features on top of generative models? - Should the stories come from professional work, or are school and personal projects acceptable? ### Part 1 — A conflict Tell me about a time you had a conflict with a coworker. What was the disagreement, what did you do about it, and how did it end? ```hint Name the decision at stake Pick a disagreement about a concrete decision with real consequences, and be ready to state the other person's position as fairly as your own. ``` ```hint The outcome is not only the relationship Prepare to say what happened to the work itself, and what you would do differently next time. ``` #### What This Part Should Cover - A specific disagreement with clear stakes, and both positions stated fairly - The candidate's own actions: understanding the other side, using evidence to decide, escalating only when warranted - A clear resolution, including committing to a decision the candidate did not favor if that is what happened - A concrete outcome for the work, and a lesson the candidate applied later ### Part 2 — Using generative AI Tell me about how you have used generative AI in your work. What did you use it for, how did you make sure the results were right, and what difference did it make? ```hint One use, in depth Choose a single task where generative AI did real work, rather than a tour of the tools you have tried. ``` ```hint Show where it fell short Be ready to describe where the output was wrong or risky, and how you noticed. ``` #### What This Part Should Cover - A concrete use case and the candidate's own role in it - How the output was verified before anyone relied on it - A measurable effect on speed, quality or scope - Judgment about limits: correctness, sensitive data, and tasks where generative AI was not the right tool ### What a Strong Answer Covers - Two distinct, specific stories, each with a situation, the candidate's task, their actions and the result - First-person ownership of decisions and actions, not a vague "we" - Results stated with numbers or clearly observable outcomes - Honest reflection, including what went wrong - A calibrated view of generative AI: neither uncritical enthusiasm nor dismissal ### Follow-up Questions - If I asked the other person in that conflict, what would they say you could have done better? - Tell me about a time a decision went against your position. How did you support it afterwards? - Describe a time AI output looked correct but was not. How did you find out? - What kinds of work would you not hand to a generative AI tool, and why?