The Feature Nobody Scheduled

“You find a 40% cold start improvement while profiling. It's not on any roadmap. How do you make it happen?”

The Situation

While profiling a critical onboarding flow, you identify that three heavyweight SDKs are initialized on the main thread at startup, adding ~600ms of cold start time on mid-range devices. You estimate a 40% cold start reduction is achievable in one week of work. This affects the majority of new user first sessions. No PM has asked for this. It's not in any sprint or roadmap.

Context

Cold start time is a known pain point but hasn't been formally prioritized this quarter
The SDK initialization code is partially owned by a platform team — you'd need their involvement
Cold start time correlates with new user retention at scale — internal estimates suggest a 40% cold start improvement could yield a measurable D7 retention lift, though exact figures vary by app category.
You estimate the 40% improvement with high confidence after a 2-hour proof-of-concept
Your team is currently in the middle of a 6-week feature sprint with no slack capacity planned

The Question

“Tell me about a time you identified and drove a significant technical improvement that wasn't formally requested or prioritized.”

Response Options

One of these is the strongest response. The others reflect common approaches with real trade-offs.

I wrote up the opportunity as a ticket, added it to the backlog, and flagged it to my manager for consideration in future planning.

I built a proof-of-concept over two days to validate the 40% estimate, wrote a one-page impact brief connecting cold start time to the retention data we already had, aligned informally with the platform team EM, and brought it to our PM with a concrete ask: one week of time tracked as a tech investment with a retention metric commitment.

I implemented the optimization during off-hours and submitted a PR, letting the quality of the work speak for itself.

I raised it in our weekly tech sync and asked if anyone wanted to pick it up as a side project.

The Debrief

Why the Best Response Works

Answer B works because it makes saying 'yes' easy. By the time the PM receives the ask, the technical risk is quantified (PoC exists), the business case is made (retention data linkage), and the cross-team dependency is already resolved informally. The PM's only decision is whether to allocate a week — all other objections have been pre-handled.

What to Avoid

The backlog ticket approach is the single most common failure mode at Staff level. Tickets require someone else to prioritize your idea. Strong Staff engineers own the prioritization journey — they don't just file the ticket, they do the work to make the ticket undeniable.

What the Interviewer Is Probing

The interviewer is measuring proactive ownership and business acumen. Can you identify impact that isn't visible in the roadmap? Can you translate an engineering metric into a business metric? Can you navigate the organizational steps needed to get unsolicited work approved?

SOAR Structure

**Situation:** Profiling revealed 600ms of main-thread SDK init at startup — 40% cold start reduction achievable. Not on roadmap. **Obstacle:** No PM ask, platform team dependency, mid-sprint capacity constraints. **Action:** 2-day PoC to validate estimate; impact brief connecting startup time to retention; informal EM alignment with platform team; specific 1-week ask to PM with metric commitment. **Result:** PM approved; shipped in 6 days; cold start reduced 38%; new user D7 retention +1.2% (within expected range).

The Learning Arc

"This taught me that unsolicited work only gets approved when the proposer has already done the approval work for everyone else. The PM doesn't have to decide anything — I've already solved the cross-team dependency and proved the impact. I now apply this as a template: find the opportunity, prove it fast, connect to a business metric, pre-align dependencies, then ask with a specific commitment."

IC Level Calibration

senior

Build the proof-of-concept, document the opportunity, bring it to your manager with a clear time estimate and impact case. Ask for explicit approval before proceeding.

staff · Primary Target

Validate the opportunity (PoC), tie the metric to business impact (retention data), pre-align the platform team EM, and present it to the PM with a specific ask and a metric commitment. Own the cross-team coordination.

principal

Identify this as a signal of a broader systemic gap (no startup time budget in place). Propose the optimization as the first entry in a startup performance budget. Make it the example that prompts a platform-wide initiative — and ensure the instrumentation lands in every team's definition of done.

Company Calibration

Amazon

LP: Invent and Simplify + Bias for Action

Google

Googleyness: Proactivity + user impact orientation