Google Cloud Professional Data Engineer

About the Professional Data Engineer exam

Google Cloud Professional Data Engineer is for people who move, store and prepare data on Google Cloud. It checks whether you can design pipelines that ingest batch and streaming data, pick the right storage for each access pattern, model a warehouse in BigQuery, keep sensitive data under control, and run the whole thing reliably without overspending. Google recommends three or more years of industry experience, including at least a year designing and managing solutions on Google Cloud. There is no prerequisite.

The questions are scenarios. A company has sensor data arriving late, a nightly job that keeps failing, a BigQuery bill that grows every month, or partners who need a copy of a dataset. You choose the service or setting that solves it. Wrong options tend to be the right product used the wrong way, a self-managed cluster where a managed service would do, or a design that copies data when it could be shared.

Exam format

• Questions: 40 to 50, multiple choice and multiple select.
• Time: 2 hours.
• Passing score: Google reports pass or fail and does not publish a cut score.
• Cost: 200 USD plus tax, online with a remote proctor or at a Pearson VUE test center. English and Japanese.
• Validity: 2 years.

Google's exam page says the exam will soon be updated for recent product renaming, so check the current exam guide for the product names it uses.

What the exam covers

The official exam guide has five sections:

• Designing data processing systems (~22%). IAM, encryption and data sovereignty, cleaning data with Dataform, Dataflow and Cloud Data Fusion, fault tolerance, and planning migrations with Database Migration Service, BigQuery Data Transfer Service and Datastream.
• Ingesting and processing the data (~25%). Choosing between Dataflow and Apache Beam, Dataproc and Spark, Pub/Sub and Kafka; streaming windows and late data; orchestration with Cloud Composer and Workflows; CI/CD for pipelines.
• Storing the data (~20%). Matching access patterns to BigQuery, BigLake, Bigtable, Spanner, Cloud SQL, AlloyDB, Firestore and Cloud Storage; warehouse modelling and how far to normalise; data lakes and governance with Dataplex.
• Preparing and using data for analysis (~15%). BI Engine and materialized views, data masking and Sensitive Data Protection, preparing data for BigQuery ML and for retrieval-augmented generation, and sharing data through Analytics Hub.
• Maintaining and automating data workloads (~18%). Ephemeral versus persistent Dataproc clusters, Composer DAGs, BigQuery editions and reservations, monitoring jobs and quotas, and multi-region resilience.

How the practice questions here are organised

There is one 10-question practice test so far, with one question on each of the main themes: choosing a store for telemetry, handling late data in a streaming pipeline, partitioning and clustering to cut query cost, nested and repeated fields in a warehouse model, ephemeral Dataproc clusters, change data capture into BigQuery, orchestrating dependent jobs, sharing datasets with partners, masking PII columns, and updating a running streaming pipeline. Every question explains why the right answer is right and why each wrong option is wrong, with a link to the Google Cloud documentation it is based on.

A study plan that works

• Take the practice test cold, then read every explanation, including those for questions you answered by guessing.
• Draw the storage decision for yourself: analytics at scale, low-latency key lookups, global transactions, relational with modest scale, and files. Most storage questions are this decision in disguise.
• Learn how Dataflow handles windows, triggers, watermarks and allowed lateness. Streaming questions turn on these details.
• Know what makes BigQuery cheaper: partitioning, clustering, avoiding SELECT *, and choosing between on-demand pricing and reservations.
• In a free-tier project, load a public dataset into BigQuery, create a partitioned table and an authorized view, and compare the bytes each query scans.

These questions are original, written from the public Google Cloud documentation and the official exam guide. They are not real exam questions, and this site is not affiliated with Google.