How to Pass DEA-C01: The Data Engineer's Guide

DEA-C01 (Data Engineer Associate) is AWS's data-pipeline certification — it tests end-to-end thinking: ingestion → storage → transformation → serving, plus the security/governance/operations wrapper around all of it. If you build data pipelines, this cert validates real skills. Plan 4 weeks at 1.5–2 hours/day; hands-on with Glue and Redshift matters more than reading.

What the exam tests

• Ingestion & transformation (34%): Kinesis, Glue ETL, DMS, batch vs streaming trade-offs
• Data store management (26%): S3 data lakes, Redshift, partitioning, storage format choices, lifecycle
• Data operations & support (22%): pipeline orchestration, monitoring, incremental processing, quality
• Data security & governance (18%): Lake Formation, encryption, access control, data quality

Week 1 — Ingestion and ETL (days 1–7)

• Days 1–2: Kinesis family — Data Streams vs Firehose vs Analytics, when each fits
• Days 3–4: Glue ETL — jobs, crawlers, the Data Catalog, job bookmarks for incremental processing
• Days 5–6: Batch ingestion — DMS, DataSync, Snow family, Transfer Family
• Day 7: practice set on ingestion questions

Week 2 — Data stores and lakes (days 8–14)

• Days 8–9: S3 data lake patterns — partitioning, columnar formats (Parquet/ORC), lifecycle tiers
• Days 10–11: Redshift — distribution/sort keys, Spectrum for lake queries, concurrency scaling, WLM
• Days 12–13: When each store wins — Redshift vs Athena vs EMR vs OpenSearch decision criteria
• Day 14: timed practice set on storage questions

Week 3 — Operations and governance (days 15–21)

• Days 15–16: Pipeline orchestration — Step Functions vs MWAA, event-driven pipelines, error handling
• Days 17–18: Data quality — Glue Data Quality, Deequ concepts, schema enforcement, DQDL
• Days 19–20: Lake Formation — permissions model, LF-tags, cross-account sharing via RAM
• Day 21: first full timed mock (65 questions, 130 minutes)

Week 4 — Security, review, and mocks (days 22–28)

• Days 22–23: Encryption and access — KMS for data services, Macie for PII discovery, IAM + LF layering
• Days 24–25: mistake review — re-do every missed question and build a weak-topic drill list
• Days 26–27: second full timed mock
• Day 28: final skim of the DEA cheat sheet + logistics

Exam traps specific to DEA

• Glue job bookmarks = incremental processing state — a frequently-tested detail
• Firehose buffers/batches/converts to Parquet automatically; Streams requires consumers
• Redshift Spectrum queries S3 WITHOUT loading — the 'occasional lake query' answer
• Lake Formation permissions layer on top of IAM — bucket policies alone bypass governance
• Kinesis on-demand vs provisioned shards — know when each is right

What 'ready' looks like

You're ready at 75%+ on timed mocks with the ability to trace a data pipeline end-to-end: where data lands, how it's transformed, who can see it, and what happens when the pipeline breaks. DEA is newer than the associate trio — its exam is respected precisely because it tests pipeline judgment, not trivia.