Prices shown are a USD base. Your local-market price is calculated at checkout in your own currency, plus any local sales tax.
What is inside
20 chapters, in the order the interviews tend to ask them:
Batch versus streaming architectures
Ingestion, connectors and change data capture
Idempotent loads and exactly-once writes
Orchestration, DAGs and scheduling
Backfills, retroactive runs and data repair
Data modelling for analytics
Warehouse versus lakehouse
Partitioning and columnar file formats
Late arriving data, windows and watermarks
Stream joins and stateful processing
Data quality, expectations and contracts
Lineage, metadata and discovery
Freshness, observability and pipeline monitoring
Transformation frameworks and query engines
Testing data pipelines
Continuous integration and delivery for data
Data mesh, domains and ownership
Real-time analytics and serving models
Cost control and data FinOps
Reproducibility, versioning and restoring data
How the answers are built
Each chapter opens with the checklist the question is really testing. A few from this pack:
Batch is a bounded read of a snapshot of data, while streaming is an unbounded sequence processed as it arrives, and both often run on the same engine.
Connectors hide source-specific details but never hide source semantics, so replication privileges, log positions and cursor behaviour still decide what arrives.
At-least-once delivery is unavoidable in practice, so idempotency is a property of the sink write rather than of the reader.
A DAG encodes dependencies, not a schedule; the scheduler decides when a task is eligible given upstream state and its own limits.
A backfill is a normal run with a different date range, so the pipeline should support it by construction rather than by hand.
Declare the grain of a fact table first, in a sentence and in a test, because every later decision follows from it.
Questions this pack covers
When should you argue against building a streaming pipeline at all?
What latency is the business actually willing to pay for here?
When do you choose log-based CDC over scheduled extracts?
What breaks when a source table column changes type?
What makes a write idempotent rather than merely safe?
Why is a natural business key not always unique?
How do you split work into DAG tasks sensibly?
When is a sensor better than a fixed schedule?
Each one is answered in full in the pack, with the reasoning written out rather than a bullet list.
About this pack
How many questions?
360, spread across 20 topics, written out in full across 236 pages.
Do the answers just list keywords?
No. Each answer is written as prose, because that is what you have to reproduce when someone asks you the question.
How do I get it?
Pay once and the PDF arrives by email immediately, with a signed download link.
Can I get it refunded?
Yes, if it is not what you expected. See the refund policy.
