As a manager, how do you decide which technical decisions to make yourself, which to delegate, and which to escalate?
leadership · Manager level · engineering-management
What the interviewer is really asking
Assesses whether the candidate has a deliberate framework for technical decision authority that grows their engineers and reserves their own involvement for genuinely high-stakes or irreversible calls, rather than over- or under-engaging by habit.
What to say
- Offer an explicit filter — reversibility and blast radius are the main axes: low-stakes/reversible decisions get fully delegated, high-stakes/irreversible or cross-org ones you own or escalate
- Describe how you delegate the decision and the authority (not just the analysis) with clear guardrails, so engineers grow real judgment rather than rubber-stamping you
- Explain how you stay informed enough to catch a bad call early without becoming the bottleneck — lightweight design reviews, escalation triggers, writing decisions down — and how you handle disagreeing with a delegated call you'd have made differently
What to avoid
- Making all the meaningful technical decisions yourself because you're 'ultimately accountable,' which caps your engineers and makes you the bottleneck
- Delegating everything with no guardrails or visibility, then being blindsided by an expensive irreversible mistake
- Overriding delegated decisions whenever they differ from your preference, which teaches the team that delegation is theater
Example answers
Strong: I use reversibility as my main filter. Anything reversible and within one service, the owning engineer decides outright — I want them building judgment, not waiting on me. Irreversible or cross-team calls (a data model that's hard to migrate, picking a vendor we'll be locked into) I either facilitate a decision on or escalate if it's above my pay grade. I made the rule explicit so the team knows when to just go and when to loop me in, and design docs only need my sign-off in the second category.
Weak: I review and approve all the significant architectural decisions personally — it's my name on the line if something goes wrong.