DMML references -- books, patterns, tools and terms.

The course-aligned reading shelf for Data Management for Machine Learning: data engineering lifecycle, data-intensive systems, dimensional modeling, pipeline practice, ML reliability and data infrastructure. Use this as a focused reference list, not as a shopping list.

Primary books from the decks

• Reis & Housley -- Fundamentals of Data Engineering -- best overall match for the course lifecycle: generation, storage, ingestion, transformation, serving, security, DataOps and architecture.
• Martin Kleppmann -- Designing Data-Intensive Applications -- data models, encoding/serialization, schema evolution, replication, consistency, services, logs and streams.
• Chen, Murphy, Parisa, Sculley & Underwood -- Reliable Machine Learning -- ML pipeline reliability, monitoring, incident response, data/model drift and production discipline.
• James Densmore -- Data Pipelines Pocket Reference -- practical pipeline design, ELT/ETL, orchestration and operational concerns.
• Ville Tuulos -- Effective Data Science Infrastructure -- data science/ML platform layers, compute, scheduling, versioning and model development.
• Kimball & Ross -- The Data Warehouse Toolkit -- dimensional modeling, facts, dimensions, star schemas, snowflake schemas and marts.

Reference map by lecture

Lecture	Read first	Use it for
CS1	DDIA + FDE	Data models, serialization, storage layout, OLTP/OLAP and ACID/BASE.
CS2	FDE + Reliable ML	Data as asset/liability, governance, reliability and ML data sensitivity.
L3	FDE + Kimball	Warehouse/lake/lakehouse, marts, star/snowflake, mesh/fabric and big-data architectures.
L4	DDIA + Densmore	Dataflow through DB/services/messages, schema evolution, ETL/ELT, batch, streaming and CDC.
L5	FDE + Tuulos + Reliable ML	Infrastructure, orchestration, serving, DataOps, CT/CD and observability.

Terms to keep straight

Term	One-line meaning
Data product	Owned, discoverable, reliable data asset with defined consumers and quality contract.
Metadata	Data about data: schema, owner, freshness, lineage, quality, access policy.
Lineage	Where data came from and which transformations produced it.
CDC	Change data capture: stream inserts/updates/deletes from a source database.
Reverse ETL	Move warehouse/model outputs back into operational tools like CRM.
DataOps	DevOps discipline applied to data pipelines: tests, versioning, monitoring and automation.
CT/CD	Continuous training and continuous deployment for ML systems.
Data swamp	Data lake without metadata, quality, governance or trust.

Tool map

Layer	Common tools	What to remember
Ingestion	Fivetran, Airbyte, Debezium, Kafka Connect	Connectors move data from DBs, APIs, SaaS and files.
Streaming	Kafka, Kinesis, Pulsar, Pub/Sub	Durable logs decouple producers and consumers; replay matters.
Storage	S3, GCS, Azure Blob, HDFS	Object stores and distributed filesystems back lakes and lakehouses.
Warehouse	Snowflake, BigQuery, Redshift, Synapse	Analytical SQL, columnar storage and scalable query execution.
Lakehouse	Delta Lake, Apache Iceberg, Apache Hudi	Table semantics, transactions, schema evolution and time travel on object storage.
Transform	dbt, Spark, Flink, Beam	SQL/model transforms, batch/stream compute and reusable transformations.
Orchestrate	Airflow, Dagster, Prefect	Dependencies, scheduling, retries, backfills and observability.
Catalog/govern	DataHub, OpenMetadata, Amundsen, Collibra	Discovery, lineage, ownership, glossary and metadata.
Quality	Great Expectations, Soda, dbt tests	Schema, null, uniqueness, range and freshness checks.
ML platform	MLflow, Feast, Kubeflow, Tecton	Experiment tracking, feature stores, training pipelines and deployment.

Useful free reads

• Data Engineering Wiki -- broad community map of data engineering terms and tools.
• Martin Fowler -- How to Move Beyond a Monolithic Data Lake to a Distributed Data Mesh -- domain ownership and data-as-a-product thinking.
• Martin Fowler -- Evolutionary Database Design -- schema change discipline and compatibility.
• Confluent -- Stream Processing series -- practical introduction to event streams and stream processing architecture.
• dbt docs -- Analytics engineering introduction -- SQL transformation, tests, lineage and documentation.
• Apache Airflow docs -- orchestration concepts and DAG scheduling.
• Delta Lake, Apache Iceberg, Apache Hudi -- lakehouse table formats.

Rest of this vault

• Slides Explained -- all five decks in plain words.
• Cheatsheet -- one-card recap.
• Question Bank -- questions only, no answers.
• Formula Sheet -- metrics and formulas.
• Book Guide -- how the books map to lectures.