Agile &amp; Scrum

How can developers support the Product Owner in backlog refinement?

Developers surface risks, sizing, and dependencies; co-create trade-offs; and split items early.

Decompose a monolith for scaled agile teams

Tests aligning architecture to team boundaries during incremental monolith decomposition. Cover: bounded contexts with isolated data and sagas, backward-compatible versioned APIs, and replacing shared libraries with duplicated code or versioned SDKs.

Describe a SAFe rule that hinders agility and how to mitigate it

Cite one rigid workflow pattern, show how it delays feedback, then propose a lighter cross-team substitute.

How would you estimate a cross-cutting initiative in PI Planning?

Tests decomposition of cross-cutting work into team enablers with visible dependencies. Good answer: teams estimate own slices in normalized points, map dependencies on the ART board, and reserve IP buffer.

What is a Scrum of Scrums purpose and what technical info is shared?

Multi-team sync for blockers, dependencies, API changes, integration risks; not a status meeting.

How would you apply Little's Law to optimize Kanban WIP limits?

Tests whether you can operationalize queueing theory in Kanban. A strong answer cites L equals lambda times W, fixes throughput, then solves for a WIP limit that yields a target cycle time.

What does a widening CFD Code Review band indicate?

Tests CFD literacy: a widening Code Review band shows WIP accumulation and a bottleneck. Great answers cite WIP limits, swarming, and policy fixes before hiring. Red flag: mistaking inventory growth for increased throughput.

Design an Upstream Kanban process for product ideas before development

Tests your grasp of pre-commitment demand shaping. A strong answer maps an option-discovery board, defines the commitment point and triage policies, and ties early filtering to reduced downstream variability.

CFD Testing band widens: what does it indicate and what experiments?

This tests flow metric literacy. A widening Testing band means arrivals exceed departures; propose experiments like smaller batches, automation, or dev-test swarming, then measure cycle time. Red flag: blaming testers or demanding headcount without data.

How would you implement Classes of Service in Kanban?

Tests whether you segment work by risk and cost of delay. A strong answer defines explicit policies, visualizes classes with color or lanes, and reserves WIP capacity per class. Red flag: using classes as simple priorities without capacity rules.

Lead Time vs Cycle Time in Kanban and measuring Cycle Time

Tests whether you distinguish customer wait time from active work. Strong answer: Lead Time is request-to-delivery with queues; Cycle Time is active start-to-finish measured from In Progress to Done. Red flag: treating them as synonyms or ignoring wait states.

Describe the initial columns for a new Kanban board and their purpose

Tests whether you understand Kanban as a flow visualization tool. A strong answer names Backlog, To Do, In Progress, and Done, explaining each as a handoff or state change. Red flag: adding too many columns upfront or conflating the board with Scrum.

Describe a framework to strategically manage tech debt during product discovery

This tests strategic debt tradeoffs under speed pressure. A strong answer classifies debt by interest, caps MVP debt with guardrails, and reserves fixed sprint capacity for repayment. Red flag: vilifying debt or deferring cleanup without triggers.

How would you measure a launched feature's success and impact?

This tests if you link code to business outcomes via agile metrics. A strong answer covers value, quality, satisfaction; names metrics like velocity or cycle time; and uses reports to track progress. Red flag: defining success purely by uptime or bug counts.

How do you differentiate an MVP from a throwaway prototype architecturally?

Distinguish by user commitment; define bounded contexts with stable interfaces; favor reversible decisions and day-one observability.

How would you implement an A/B test for a redesigned dashboard?

Tests experiment rigor: randomize stable user buckets, pick one primary metric and guardrails like latency, pre-calculate sample size and runtime with a minimum detectable effect, and validate via backend logging.

Describe your role as an engineer in story refinement

Mention feasibility probes, acceptance criteria checks, and splitting for forecast clarity.

What user story details reveal the customer problem?

This tests if you see stories as problem placeholders, not specs. A strong answer asks for user role, action, 'so that' value, and confirmation criteria while demanding conversation. Red flag: listing technical tasks without mentioning the customer problem.

When should cross-cutting concerns use a shared versus duplicated solution?

Weigh change frequency and blast radius; decide via fitness function or ADR, not default shared code.