Walk Through a Project, a Project That Went Wrong, and How You Use AI

A culture-focused conversation built around three open questions. Answer each one as you would to an interviewer who keeps asking "why" and "what exactly did you do". ### Clarifying Questions - Should the project in the first question be your most impactful one, your most recent one, or one that best shows how you work? - For the failed project, does "went wrong" include projects that shipped late or with problems, or only ones that were cancelled? - Does "how do you use AI" refer to your day-to-day engineering work, to AI features you have built, or both? ### Part 1 — Tell me about a project Walk the interviewer through a project you worked on: what it was for, what you personally did, and how it turned out. ```hint Pick for depth, not size Choose a project where you can explain your own decisions and their trade-offs in detail, even if it was not the largest one you touched. ``` #### What This Part Should Cover - The problem and why it mattered, in terms a non-specialist can follow - Your own role and decisions, clearly separated from the team's work - A result backed by something concrete, such as a metric, a user outcome or a date that was met - What you would do differently now ### Part 2 — Tell me about a project that went wrong Describe a project that did not go as planned: what happened, what part you played in it, and what you did about it. ```hint Own a real share of it Pick a failure where some of the cause was your own decision or omission, and be ready to describe what you changed in your behavior afterwards. ``` #### What This Part Should Cover - What went wrong and the actual impact, stated without minimizing it - Your share of the cause, with no blaming of others - How you responded while it was going wrong, and how it was recovered or closed - A lasting change in how you work, with evidence that you applied it later ### Part 3 — How do you use AI? Explain how you use AI tools in your work today. ```hint Show judgment, not enthusiasm Describe concrete tasks where the tools help you, how you check what they produce, and where you deliberately do not rely on them. ``` #### What This Part Should Cover - Specific, repeatable uses rather than general statements - How you verify AI output before it reaches code review, users or production - Limits you respect, such as confidential data, licensing or tasks where the tools are unreliable - How the tools have changed your speed or quality, with an honest example ### What a Strong Answer Covers - Real stories with specific actions, decisions and results, told in a clear situation-task-action-result order - A consistent picture across all three answers of how you make decisions, collaborate and learn - Honest self-assessment, especially in the failure story - Concise answers that leave room for follow-up questions ### Follow-up Questions - In the first project, what was the hardest technical or people disagreement, and how was it resolved? - In the project that went wrong, when was the first warning sign, and why was it not acted on earlier? - Tell me about a time an AI tool gave you a wrong answer that looked right. How did you catch it? - How would your teammates describe the way you work, and what would they want you to improve?