I’ve gotten lots of mileage out of “How do big decisions get made at this company?” And you want to turn it into a conversation – probe their answers, ask for examples, etc.
You learn about how each interviewer experiences the decision-making process. How are the decisions communicated? Who decides? Do initiatives steamroll people and teams? Are people able to filter up suggestions or start their own initiatives? Can decisions get made team-by-team, or do they have to be made for the whole company? Do they change their mind when they get new evidence? Do people change their mind too much? Do they have trouble saying “no”? If they want to tell you “no,” do they actually say it? Are they too afraid to make decisions that can change the culture, and just kinda drift?
As an IC engineer, this is what I have the least control over (and can find the most frustrating). So I want to hear exactly how broad change happens. It helps me imagine how it’ll feel to spend four years in the environment.
This is the kind of question that you need to ask everyone. You won’t get a good answer from one person. You want to ask a few people and see if their stories line up.
– What’s the roadmap for this year? This gives me a lot of insight on what I could be working on, and also if the company is a bit clueless about their direction.
– What would I do in the position I’m being interviewed for? This completes the picture, similar to the previous one, but different perspective.
If I’m working on a feature and discover technical debt that will make it more difficult to implement, how do I decide whether to focus on that debt or the feature? Can you give an example?
The reason interviewers like to ask for examples is that it’s easy to bullshit when speaking abstractly, but people are less likely to lie to your face. Use that to your advantage
Questions about tests. I have found consistently that when an interviewer does not know what the test coverage of their codebase is, more likely than not they don’t care about such trivialities.
No matter how respectable the company might look on the outside, the codebase is an incorrigible mess built by cowboys and they will expect you to maintain it.
I think a good tech one that is a bit opinionated is: Does a project(s) build, and run (fully, without problems) in 5 or less steps?