What should a candidate ask a director to learn whether reliability work is actually funded?

## Values are free; funding is not The difference between an organization that cares about engineering quality and one that says it does is visible in exactly one place: whether non-feature work has ever displaced feature work in a planning cycle. That trade happens at the level of the leader who owns several teams and a budget, which is why this question belongs in the skip-level round and not with the engineers, who can tell you the pain but not the decision. ## The question that works `Which piece of non-feature engineering work was funded in the last planning cycle, and what did it displace?` Every clause is doing work. *Non-feature* covers reliability, migrations, deleting an old path, paying down an interest-bearing design mistake. *Funded* asks for people and time rather than permission. *Last planning cycle* pins it to something that actually happened instead of something intended. And *what did it displace* is the clause that cannot be answered with goodwill, because if nothing was displaced, nothing was traded, and the work was fitted into the cracks. ## Reading the answers **Specific and costed.** A leader names the workstream, says roughly who was on it — a range of two to four engineers for a cycle is a normal shape — and names the commitment that slipped. This is the answer you want. Note also whether they can say what it bought: fewer incidents on a service, a dependency removed, an on-call rotation that stopped waking people. **Specific but uncosted.** A named workstream with no displaced commitment usually means the team absorbed it on top of the roadmap. That can be sustainable once; as a pattern it is how teams end up quietly over-committed. **`We trust teams to fit it in.`** This sounds like autonomy and often functions as its opposite: unprotected work loses to committed work whenever a quarter gets tight, which is most quarters. Probe once by asking when a team last used that latitude and what happened to their other commitments. **`Quality is one of our values.`** A non-answer. Reframe rather than repeat: ask what would have to be true for a reliability effort to get a named owner and time next cycle. That is answerable without disclosing plans, and the mechanism it reveals is the actual answer. ## The same test applied to career growth Investment in people has the same shape. The weak version of the question asks whether there is a path to senior; the answer is always yes. The strong version asks for cases: where did the last few people promoted into senior roles across this group come from, were they internal or hired in, and what work made the case. A leader at this level can answer that from memory if it happens, and cannot invent it convincingly if it does not. If every senior on the org chart you are sketching arrived from outside, the growth story for the role you are considering is a hiring story, not a development one. Promotion frameworks and their timing vary a great deal between companies, so treat what you hear as a description of this group rather than a general rule. ## Where the answers land on the sketch Add two annotations to your org chart: which teams got non-feature capacity last cycle, and where the internally promoted seniors sit. Both cluster. Teams attached to the highest-ranked bets usually get more of both, because visibility funds things. If the team you are considering is attached to the third of three bets and has never displaced a commitment for engineering work, you now know two concrete things about what your next couple of years would look like, from questions that took four minutes to ask. ## Tone Ask it as someone planning to do the work, not as an auditor: you want to know how a proposal for this kind of work would get made and funded, because you expect to make one. That framing is true, it is flattering in an honest way, and it usually produces the specific answer, because leaders like talking about how decisions actually get made in their group.