Interview prompt
Problem context
Skills being evaluated
Use the sequence below to surface constraints, choose boundaries, test failure behavior, and defend trade-offs. Concrete numbers are interview assumptions, not claims about a real production system.
Clarify the decision
- Reconfirm the original business outcomes, current constraints, decision rights, sunk costs, remaining migration, and what happens if the program stops. Interview teams and inspect production evidence.
Establish scale assumptions
- Measure traffic and capability migrated, dependency and bridge count, duplicated cost, incident impact, team capacity, and critical-path milestones. Replace percent-complete estimates with outcome cohorts.
Functional and non-functional requirements
- Restore one source of truth, narrow scope to value, assign end-to-end owners, freeze new divergence, and create stop or recommit gates. Status must reveal risk and uncertainty.
High-level architecture
- Define the minimum viable target boundary, authoritative migration units, bridge inventory, and compatibility plan. Form a small empowered program team while domain teams own specific cutovers and decommissioning.
Data model and flow
- A program ledger links each cohort to source, target, dependencies, data authority, validation, rollback, owner, and removal. Production metrics feed status automatically where possible.
Consistency and transaction boundaries
- Stop adding features to both systems; choose one authority per capability. Temporary synchronization has reconciliation and an expiry decision, not vague eventual retirement.
Failure modes and recovery
- Protect current customers while testing the hardest remaining cohort early. If a core assumption fails, reduce scope or stop rather than moving more easy traffic to preserve optics.
Security and privacy
- Audit temporary credentials, data copies, exceptions, and unsupported bridges accumulated during delay. Migration pressure does not justify bypassing controls.
Observability and SLOs
- Report business capability moved, traffic, reliability, unit cost, bridge retirement, blocked decisions, and forecast confidence. Publish red status when decision evidence is absent.
Capacity and cost
- Compare finish, narrow, coexist, and stop scenarios including future dual-run cost. Reallocate teams from activity that does not retire a constraint.
Alternatives and trade-offs
- Replanning delays visible output but restores a credible path; continuing current execution protects short-term optics and compounds debt. Run a bounded reset with an executive decision date.
Evolution and migration
- Freeze, audit, test the hardest assumption, choose scope, deliver one end-to-end value slice including decommissioning, then scale only if outcomes improve.
What Staff and Principal candidates should emphasize
- Principal candidates are willing to stop or narrow a prestigious program. They restore truth, decision rights, and value while treating teams fairly and removing incentives for green theater.
Decision trade-offs
Program path
Option A
Continue to preserve momentum
Option B
Pause for a bounded evidence and scope reset
Recommendation:Pause new expansion briefly, protect production, test the hardest assumptions, and make an explicit finish, narrow, coexist, or stop decision.
Milestones
Option A
Track components and tasks delivered
Option B
Track business cohorts migrated and legacy removed
Recommendation:Use end-to-end cohorts with decommissioned old behavior; component completion can increase complexity without producing value.
Follow-up interview questions
- 01How do you tell executives the program may need to stop?
- 02Which cohort should be attempted next?
- 03How do you freeze divergence from new legacy features?
- 04Who owns removing temporary bridges?
Common weak answers and mistakes
- 01Adding more coordination meetings without changing scope or authority.
- 02Continuing easy migrations while the hardest assumption remains untested.
- 03Reporting task completion as architecture value.
- 04Treating sunk cost and reputational concern as technical requirements.
Interviewer evaluation rubric
Replans dates or adds staff without examining outcomes, assumptions, authority, and stop options.
Audits reality, narrows scope, assigns owners, tests hard cohorts, and reports outcome milestones.
Compares strategic options, freezes divergence, retires bridges, restores trust, and changes incentives and governance.
Makes an evidence-based stop or recommit decision that preserves organizational learning and reallocates capital without blame.