Android Developer DBMS Concepts Interview Guide

At Read Committed, each statement receives a new snapshot, so later statements can see commits made in between. Higher isolation changes the guarantees and may require retrying serialization failures. Choose isolation from the business invariant, not from an assumption that every transaction freezes all reads.

Check your understanding: Interleave two sessions and document the snapshots each statement sees.

Common trap: Treating transaction atomicity and isolation as identical.

Concept reference: PostgreSQL: transaction isolation

An index is an option, not a command. The planner considers selectivity, statistics, table size and access cost. Read estimated and actual rows, filters and scan types before changing indexes. EXPLAIN ANALYZE executes the statement, so use safe data and transaction handling for writes.

Check your understanding: Compare estimated and actual row counts on representative data.

Common trap: Forcing an index without measuring the full query.

Concept reference: PostgreSQL: reading EXPLAIN plans

Check the grain of each table and whether the join creates several rows for one order. Aggregate at the intended business grain before joining, or use the correct relationship and key. Reconcile totals using a tiny fixture containing multiple child rows and an unmatched parent.

Check your understanding: Use one order with two child records and verify the total remains correct.

Common trap: Using DISTINCT as a universal repair for a wrong join.

Concept reference: PostgreSQL: reading EXPLAIN plans

The Activity lifecycle is not a durable data boundary. Separate UI state from the component instance and keep a defined source of truth in the data layer. Distinguish configuration changes from process death; state that must survive longer needs an appropriate saved or persistent representation.

Check your understanding: Rotate the device and separately recreate the process, checking the intended state.

Common trap: Assuming a ViewModel is durable storage across process death.

Concept reference: Android: app architecture

Expose state from a data layer with an explicit source of truth, freshness and error model. Decide which actions can be queued and how they reconcile later. Keep network and storage details out of the UI component, and test a reconnect that returns older or conflicting data.

Check your understanding: Disconnect during an edit and verify the documented reconciliation behavior.

Common trap: Replacing useful cached data with a generic error screen.

Concept reference: Android: app architecture

A scope connects task lifetime, cancellation and failure propagation. Work started without a deliberate owner can outlive the screen or operation that needs it. Choose an appropriate lifecycle or application boundary and ensure cancellation does not leave half-applied local state.

Check your understanding: Cancel the owner and verify child work finishes or cancels as designed.

Common trap: Launching every task in an unrestricted global scope.

Concept reference: Kotlin: coroutine fundamentals

First, identify technical constraints and define measurable service objectives. Next, trace a user action through lifecycle, local state, network and restored UI. Contrast architectural trade-offs across write cost, read latency, consistency and operational complexity, explicitly mitigate the risk of bad data is cached and amplified across downstream consumers, and confirm system stability using semantic validation results, quarantine records and a replay verification.

Common trap: Reaching for a specific library or framework before defining constraints, failure envelopes, and automated verification criteria.

Prioritise the failure mode exhibiting the highest user blast radius and lowest observability. Formulate an explicit containment boundary, implement idempotent retries with jitter, and establish an automated rollback threshold. Verify resilience through load-test results, retry counts and a measured cost estimate.

Common trap: Relying on passive monitoring dashboards without defining explicit error-budget alerts, rollback triggers, and verified recovery procedures.

Structure the response using Context-Decision-Tradeoff-Result: articulate the business and technical constraints, compare viable alternatives, explain the implementation (trace a user action through lifecycle, local state, network and restored UI), and document the accepted trade-off. Provide concrete proof: an explain plan, concurrency test and recovery verification.

Common trap: Speaking only in high-level abstractions or team accomplishments without detailing your direct implementation decisions, trade-offs, and measured results.

Formulate a falsifiable hypothesis from observable telemetry before altering configurations. Then compare lifecycle events, persisted state, network retries and device-specific traces. Isolate the defect to the smallest reproducible boundary, validate root cause with evidence, and confirm full resolution using a compatibility test matrix and a staged rollout metric.

Common trap: Applying speculative fixes or restarting services blindly without establishing an observable signal connected to a falsifiable hypothesis.

Produce a constrained-device test matrix with offline recovery evidence. Document baseline assumptions, technical mechanism (trace a user action through lifecycle, local state, network and restored UI), rejected alternatives, bounded failure envelopes, and deterministic pass criteria. Supply reproducible verification via a trace, a minimal reproduction and a regression test.

Common trap: Presenting architecture diagrams or slides lacking automated unit/integration tests, observable metrics, or automated rollback configurations.