Everything in Product Management, page 21
How do duplicate events bias COUNT(*) and daily login reports?
Tests idempotency in streaming analytics. COUNT(*) overcounts; fix with unique event ID dedup via idempotent writes or COUNT(DISTINCT id), plus daily partition reconciliation. Red flag: SELECT DISTINCT * without a stable key or no reporting safeguard.
Why is star schema preferred over 3NF for analytics?
Tests your grasp of the read-performance trade-off in analytical schemas. A great answer names fact and dimension tables, emphasizes fewer joins for aggregations, and cites simpler SQL and faster query plans.
Slow dashboard querying a large fact table: first three checks?
This tests systematic diagnosis of fact-table query latency. A strong answer checks the execution plan and indexing, evaluates partitioning and data model fit, and inspects caching or pre-aggregation.
Why prefer median and p95 over mean for API latency?
This tests statistical intuition for skewed distributions. A strong answer notes that median captures typical experience, p95 captures tail suffering, and mean hides outliers. A red flag is claiming mean alone is sufficient.
Design a scalable data governance framework balancing autonomy and control
Self-serve platform with domain products, auto-catalog, schema contracts, and policy-as-code access in CI/CD.
How do you root-cause bad data across microservices and Spark?
This tests structured debugging and observability for distributed pipelines. A strong answer isolates the break via lineage, validates schema and freshness per stage, and compares microservice outputs to Spark inputs.
How would you build a Customer golden record across fragmented systems?
This tests master data management discipline for distributed, conflicting records. A strong response covers identity resolution, survivorship rules, merge architecture, lineage, and feedback loops.
Design a CDC pipeline that handles schema evolution gracefully
Tests designing resilient CDC pipelines against schema drift. A strong answer covers schema registries with versioning, backward-compatible serialization, and automated compatibility checks.
How would you implement data lineage for microservices analytics?
Tests end-to-end provenance tracking, not just docs. Strong answers cover automated metadata capture at service boundaries, a central catalog such as DataHub or OpenLineage, and column-level tracing.
What is a data schema and why enforce it at ingestion?
Tests schemas as contracts and ingestion validation as a quality gate. Strong answers cite blueprints with constraints, fail-fast ingestion catching type errors upstream, and downstream trust. Red flag: treating schemas as optional docs affecting only storage.
What validation checks would you implement for a user-submitted email field?
Tests whether you separate syntax validation from deliverability and know practical ingestion guards. A strong answer covers RFC-aware syntax, domain checks, normalization, and deduplication.
How do you handle 10% nulls in a key numerical column?
Check MCAR/MAR/MNAR; prefer median or mean imputation; add an "is_missing" indicator.
Design a data quality framework from source to consumption
This tests full-lifecycle data architecture. Strong answers define ownership first, then schema contracts at ingestion, profiling and anomaly detection in CI/CD, column-level lineage, and KPI-linked scorecards. Red flag: tools before ownership or RACI.
Diagnose out-of-memory errors in a growing Spark job
Inspect plans for skewed partitions and joins; tune shuffle partitions, batch sizes, and join strategies; leverage AQE and statistics.
What is data partitioning in a cloud data warehouse?
Tests physical data layout and cost/performance tradeoffs. Strong answers define time-based or integer-range partitioning, explain partition pruning avoids full scans, and warn against high-cardinality keys.
How would you handle late-arriving data in a streaming analytics pipeline?
Tests understanding of event-time processing, watermarks, and windowing for correctness. A strong answer covers watermarks with allowed lateness, event-time triggers, and idempotent updates to sinks.
Describe star and snowflake schemas and their trade-offs
Star schemas flatten dimensions for fast joins; snowflake schemas normalize hierarchies to reduce redundancy but add joins.
Explain data warehouse purpose and how it differs from OLTP
This tests whether you know the OLTP versus analytics split. A great answer contrasts OLTP row-level writes and normalized schemas with warehouse denormalized schemas and BI reads. A red flag is calling a warehouse just a bigger OLTP database.
Outline an automated pipeline to load a daily CSV into a database
Event trigger on drop, schema validation, idempotent load, and observability.
Differences between ETL and ELT, and when to choose each
ETL transforms before loading for structured data; ELT loads raw first and transforms in the warehouse for scale.
We are hiring for this. Every open role lists the topics its interview covers, so you can prepare for the real thing rather than guessing.
See open roles