Google Cloud Professional Data Engineer Exam Prep

What's inside

1. Level 1: The Exam, Its Guide and Best-Answer QuestionsFree
   [Exam orientation] You learn the exam's facts and the habits that answer its scenario questions well: the exam guide belongs to Google, is free on Google's certification page and is not reproduced in this course; the standard exam lasts 2 hours, has 40-50 multiple-choice and multiple-select questions and costs $200; this course is independent of Google, is not endorsed by Google and cannot promise a pass; and Google is renaming products, so the course uses current names and notes the former ones.
2. Level 2: Cloud IAM and Organization Policies for DataFree
   [Section 1: Designing data processing systems — 1.1 Designing for security and compliance] You learn how Cloud IAM roles, service accounts and organization policies control who can reach data in BigQuery, Cloud Storage and pipelines, and how least privilege appears in exam scenarios.
3. Level 3: Encryption and Key Management
   [Section 1: Designing data processing systems — 1.1 Designing for security and compliance] You learn how Google Cloud encrypts data at rest and in transit, and when a requirement calls for Google-managed keys, customer-managed encryption keys (CMEK) in Cloud KMS, or keys held outside Google with Cloud External Key Manager.
4. Level 4: Handling PII with Sensitive Data Protection
   [Section 1: Designing data processing systems — 1.1 Designing for security and compliance] You learn strategies for personally identifiable information: finding it with Sensitive Data Protection (formerly Cloud DLP) discovery and inspection, and protecting it with masking, tokenization, format-preserving encryption and bucketing.
5. Level 5: Data Sovereignty and Regulatory Compliance
   [Section 1: Designing data processing systems — 1.1 Designing for security and compliance] You learn the regional rules for data access and storage: choosing regions and multi-regions, enforcing location with organization policies, and how legal requirements such as GDPR, HIPAA and data residency laws shape a design.
6. Level 6: Governed Project, Dataset and Table Design
   [Section 1: Designing data processing systems — 1.1 Designing for security and compliance] You learn how to lay out projects, BigQuery datasets and tables for good data governance, and how to separate development and production environments.
7. Level 7: Preparing and Cleaning Data Reliably
   [Section 1: Designing data processing systems — 1.2 Designing for reliability and fidelity] You learn to choose a data preparation and cleaning approach (Dataform SQL workflows, Dataflow, Cloud Data Fusion, or prompting LLMs for query generation) and to judge what each gives you in reliability.
8. Level 8: Data Validation and Pipeline Monitoring Design
   [Section 1: Designing data processing systems — 1.2 Designing for reliability and fidelity] You learn to design data validation checks and to build monitoring and orchestration into a pipeline from the start, so bad or missing data is caught before consumers see it.
9. Level 9: ACID, Availability and Disaster Recovery
   [Section 1: Designing data processing systems — 1.2 Designing for reliability and fidelity] You learn to make design decisions about ACID compliance versus availability, and to plan disaster recovery and fault tolerance using RPO and RTO targets.
10. Level 10: Flexibility, Portability and Data Discovery
    [Section 1: Designing data processing systems — 1.3 Designing for flexibility and portability] You learn to map current and future business requirements to an architecture, to design for data and application portability across clouds and residency rules, and to plan data staging, cataloging, profiling and discovery with Dataplex (being renamed Knowledge Catalog).
11. Level 11: Planning Data Migrations to Google Cloud
    [Section 1: Designing data processing systems — 1.4 Designing data migrations] You learn to analyse stakeholders, users, processes and current technology, and to plan migration and validation with BigQuery Data Transfer Service, Database Migration Service, Transfer Appliance, Storage Transfer Service, Google Cloud networking and Datastream.
12. Level 12: Planning Pipelines: Sources, Sinks and Logic
    [Section 2: Ingesting and processing the data — 2.1 Planning the data pipelines] You learn to define data sources and sinks, and to separate transformation logic from orchestration logic, before choosing any service.
13. Level 13: Pipeline Networking and Encryption
    [Section 2: Ingesting and processing the data — 2.1 Planning the data pipelines] You learn the networking fundamentals and data encryption choices that pipelines depend on: VPCs and subnets, Shared VPC, Private Google Access, Private Service Connect, firewall rules for workers, and CMEK on pipeline resources.
14. Level 14: Streaming Ingestion with Pub/Sub and Kafka
    [Section 2: Ingesting and processing the data — 2.2 Building the pipelines] You learn how Pub/Sub topics, subscriptions, ordering, dead-letter topics and replay work, and when Apache Kafka (including Google Cloud Managed Service for Apache Kafka) is the better choice.
15. Level 15: Apache Beam and Dataflow Batch Pipelines
    [Section 2: Ingesting and processing the data — 2.2 Building the pipelines] You learn the Apache Beam programming model (PCollections, PTransforms, ParDo, GroupByKey, side inputs) and run batch transformations on Dataflow, including templates and autoscaling.
16. Level 16: Streaming Transformations: Windows and Late Data
    [Section 2: Ingesting and processing the data — 2.2 Building the pipelines] You learn streaming processing logic in Beam and Dataflow: event time versus processing time, fixed, sliding and session windows, watermarks, triggers, allowed lateness and handling late arriving data.
17. Level 17: Dataproc, Spark and the Hadoop Ecosystem
    [Section 2: Ingesting and processing the data — 2.2 Building the pipelines] You learn to run Apache Spark and Hadoop ecosystem workloads on Dataproc (now Managed Service for Apache Spark), including serverless batches, and when to pick it over Dataflow.
18. Level 18: Cleansing with Cloud Data Fusion and BigQuery
    [Section 2: Ingesting and processing the data — 2.2 Building the pipelines] You learn data cleansing and batch processing logic with Cloud Data Fusion's visual pipelines and Wrangler, and with in-warehouse ELT in BigQuery SQL, including deduplication and MERGE.
19. Level 19: AI Data Enrichment in Pipelines
    [Section 2: Ingesting and processing the data — 2.2 Building the pipelines] You learn AI data enrichment: adding sentiment, entities, classifications or generated text to records with BigQuery's generative AI functions and remote models, and with Dataflow's RunInference transform.
20. Level 20: Data Acquisition, Import and New Sources
    [Section 2: Ingesting and processing the data — 2.2 Building the pipelines] You learn the ways to acquire and import data into Google Cloud, and to integrate new data sources: BigQuery load jobs, the Storage Write API, external tables, Datastream change data capture into BigQuery, and BigQuery Data Transfer Service connectors.
21. Level 21: Operationalizing Pipelines with Cloud Composer
    [Ingesting and processing the data] Serves 2.3 'Job automation and orchestration'. You learn when Cloud Composer (managed Apache Airflow) is the right orchestrator, how a Composer environment is built, and how to use the Google Cloud operators and sensors to trigger and chain BigQuery, Dataflow and Dataproc jobs. Writing DAGs for repeatability is left to Section 5.2.
22. Level 22: Workflows, Cloud Scheduler and Event-Driven Triggers
    [Ingesting and processing the data] Serves 2.3 'Job automation and orchestration (e.g., Cloud Composer and Workflows)'. You compare Workflows, a serverless step-based orchestrator that calls APIs, with Cloud Composer, and use Cloud Scheduler and Eventarc to start pipelines on a schedule or when an event such as a new file arrives.
23. Level 23: CI/CD for Data Pipelines
    [Ingesting and processing the data] Serves 2.3 'CI/CD (Continuous Integration and Continuous Deployment)'. You build a CI/CD path that tests and promotes Dataflow Flex Templates, Dataform SQL and Composer DAGs from development to production, using Cloud Build, Artifact Registry and infrastructure as code.
24. Level 24: Analyzing Data Access Patterns to Pick Storage
    [Storing the data] Serves 3.1 'Analyzing data access patterns' and 'Choosing managed services'. You build the storage decision tree the exam expects, sorting workloads by structure, read/write ratio, latency, consistency, scale and query style into analytical, relational, wide-column, document, in-memory and object storage.
25. Level 25: Relational Choices: Cloud SQL, AlloyDB, Spanner
    [Storing the data] Serves 3.1 'Choosing managed services' for relational workloads. You contrast Cloud SQL, AlloyDB for PostgreSQL and Spanner on scale, availability, global consistency, analytical acceleration and cost, so you can pick the right one when a scenario needs ACID transactions.
26. Level 26: Bigtable, Firestore and Memorystore
    [Storing the data] Serves 3.1 'Choosing managed services' for non-relational workloads. You learn Bigtable's wide-column model and row-key design for time-series and IoT data, Firestore for document and mobile/web app data, and Memorystore for in-memory caching.
27. Level 27: Storage Costs, Performance and Data Lifecycle
    [Storing the data] Serves 3.1 'Planning for storage costs and performance' and 'Lifecycle management of data'. You cover Cloud Storage classes, Autoclass and lifecycle rules, retention policies, and BigQuery storage costs: long-term storage, physical vs. logical billing, table and partition expiration, and time travel.
28. Level 28: Warehouse Data Modeling and Normalization
    [Storing the data] Serves 3.2 'Designing the data model', 'Deciding the degree of data normalization' and 'Mapping business requirements'. You model a BigQuery warehouse with star schemas, denormalized wide tables, and nested and repeated fields (STRUCT/ARRAY), and weigh join cost against update complexity.
29. Level 29: Partitioning and Clustering for Access Patterns
    [Storing the data] Serves 3.2 'Defining architecture to support data access patterns'. You design BigQuery tables with time-unit, ingestion-time or integer-range partitioning and up to four clustering columns, and check that partition pruning and block pruning actually cut the bytes scanned.
30. Level 30: Managing a Data Lake with BigLake
    [Storing the data] Serves 3.3 'Managing the lake (configuring data discovery, access, and cost controls)'. You build a Cloud Storage data lake governed through BigLake tables and Apache Iceberg, with connection-based access delegation, fine-grained security on open-format files, discovery, and controls on storage and query cost.
31. Level 31: Processing and Monitoring the Data Lake
    [Storing the data] Serves 3.3 'Processing data' and 'Monitoring the data lake'. You process lake data in place with BigQuery, Dataproc Serverless for Spark and Dataflow, and monitor the lake's health, freshness, quality, usage and cost with data profiling, data quality scans, audit logs and storage insights.
32. Level 32: Data Platform and Federated Governance with Dataplex
    [Storing the data] Serves 3.4 'Designing for a data platform' and 'Building a federated governance model'. You design a platform with Dataplex Universal Catalog (this covers the guide's 'Dataplex' and 'Dataplex Catalog'; Data Catalog was folded into it), BigQuery and Cloud Storage: lakes, zones and assets, aspects and glossaries, lineage, and data products owned by domain teams under central policy.
33. Level 33: Connecting BI Tools and Precalculating Fields
    [Preparing and using data for analysis] Serves 4.1 'Connecting to tools' and 'Precalculating fields'. You connect Looker (LookML semantic layer), Looker Studio and Connected Sheets to BigQuery, and decide which metrics to precompute in scheduled queries, Dataform tables or LookML derived tables so that BI queries do not recompute them.
34. Level 34: BI Engine and Materialized Views
    [Preparing and using data for analysis] Serves 4.1 'BigQuery features for business intelligence (e.g., BI Engine, materialized views)'. You speed up dashboards with BI Engine in-memory reservations and with materialized views (automatic refresh, smart tuning, limits), and contrast them with logical views and cached results.
35. Level 35: Troubleshooting Poor-Performing Queries
    [Preparing and using data for analysis] Serves 4.1 'Troubleshooting poor performing queries'. You read the BigQuery query execution graph and INFORMATION_SCHEMA.JOBS to find slow stages, skew, large shuffles and spills, and fix common anti-patterns.
36. Level 36: Securing Analytics: Masking, IAM and Cloud DLP
    [Preparing and using data for analysis] Serves 4.1 'Security, data masking, Identity and Access Management (IAM), and Cloud Data Loss Prevention (Cloud DLP)'. You protect BI data with policy tags for column-level security, dynamic data masking, row-level access policies and authorized views, and with Sensitive Data Protection (formerly Cloud DLP) for discovery and de-identification.
37. Level 37: Feature Prep, Training and Serving with BigQuery ML
    [Preparing and using data for analysis] Serves 4.2 'Preparing data for feature engineering, training and serving machine learning models (e.g., BigQueryML)'. You engineer features in SQL, use the TRANSFORM clause so that training and serving apply the same preprocessing, train and evaluate models, and serve predictions in BigQuery or through Gemini Enterprise Agent Platform (formerly Vertex AI).
38. Level 38: Preparing Unstructured Data for Embeddings and RAG
    [Preparing and using data for analysis] Serves 4.2 'Preparing unstructured data for embeddings and retrieval-augmented generation (RAG)'. You turn documents and images in Cloud Storage into object tables, chunk and parse them, generate embeddings with remote models, and serve retrieval with vector indexes and VECTOR_SEARCH.
39. Level 39: Sharing Data with BigQuery Sharing (Analytics Hub)
    [Preparing and using data for analysis] Serves 4.3 'Sharing data'. You define sharing rules, publish datasets as listings in BigQuery sharing (formerly Analytics Hub) exchanges for zero-copy subscriptions, use data clean rooms and egress controls, and publish reports and visualizations to their audiences.
40. Level 40: Minimizing Cost While Protecting Critical Workloads
    [Maintaining and automating data workloads] Serves 5.1 'Minimizing costs per required business need for data' and 'Ensuring that enough resources are available for business-critical data processes'. You cut spend with maximum bytes billed, custom quotas, Dataflow FlexRS and autoscaling limits, and Spot VMs, while keeping capacity guaranteed for critical jobs.
41. Level 41: Persistent vs. Job-Based Spark Clusters
    [Maintaining and automating data workloads] Serves 5.1 Optimizing resources: you learn when to keep a persistent Dataproc cluster and when to use a job-based one (an ephemeral cluster created for a job, or the serverless deployment). Note the product names: Google now calls Dataproc on Compute Engine and Serverless for Apache Spark (formerly Dataproc Serverless) 'Managed Service for Apache Spark', but the exam guide still says Dataproc.
42. Level 42: Repeatable Cloud Composer DAGs
    [Maintaining and automating data workloads] Serves 5.2 Designing automation and repeatability: you learn to write Cloud Composer (Apache Airflow) DAGs that are idempotent and can be backfilled, so any run can be repeated safely and gives the same result.
43. Level 43: BigQuery Editions, Reservations and Batch Jobs
    [Maintaining and automating data workloads] Serves 5.3 Organizing workloads based on business requirements: you learn capacity management with BigQuery Editions, reservations, assignments and autoscaling slots, and when to run queries as interactive or batch jobs.
44. Level 44: Observability for Data Processes
    [Maintaining and automating data workloads] Serves 5.4 Monitoring and troubleshooting processes, observability and monitoring planned usage: you learn to watch pipelines and the warehouse with Cloud Monitoring, Cloud Logging and the BigQuery admin panel.
45. Level 45: Troubleshooting Errors, Billing and Quotas
    [Maintaining and automating data workloads] Serves 5.4 troubleshooting error messages, billing issues and quotas, and managing workloads (jobs, queries and compute capacity through reservations): you learn to diagnose the failures and cost surprises the exam describes in its scenarios.
46. Level 46: Fault Tolerance, Restarts and Multi-Region Jobs
    [Maintaining and automating data workloads] Serves 5.5 Maintaining awareness of failures: you learn to design pipelines that tolerate failures and restart cleanly, and to run jobs across multiple zones or regions.
47. Level 47: Corruption, Missing Data, Replication, Failover
    [Maintaining and automating data workloads] Serves 5.5 preparing for data corruption and missing data, and data replication and failover (e.g., Cloud SQL, Redis clusters): you learn to recover bad or lost data and to plan failover for stateful stores.
48. Level 48: Timed Mock Exam: 50 Questions in 120 Minutes Timed mock
    A timed 50-question, 120-minute mock exam with a 70% pass mark, made of original scenario-based best-answer questions across all five sections of the Professional Data Engineer exam guide, in proportion to their weights.

Access

The first 2 levels are free with a free account. Every level, the podcast edition and the AI tutor come with All Access at £4.99/month or any Creator plan — see pricing.

Start free →