Entry-level certification validating core Google Cloud concepts, services, and digital transformation capabilities.
Why cloud adoption, Digital transformation concepts, Cloud vs on-premises infrastructure, Total Cost of Ownership (TCO)
Database & Storage Options, BigQuery, Data Pipelines, Looker & Data Visualization
AI/ML Fundamentals, Vertex AI, Pre-trained APIs, Generative AI & Gemini
Compute Engine & GKE, App Engine & Cloud Run, Serverless Computing, Containers & Microservices, API Management
IAM & Zero Trust, Data Encryption, Network Security, Compliance & Privacy
Financial governance & cost management, Monitoring & SRE, Cloud Operations Suite, Organizational change management
Try real exam-style questions from the free sample set. Each answer comes with a full explanation.
A recruitment platform is introducing a system that uses AI to automatically evaluate applicants. Management determined that they need to review whether the AI disadvantages specific groups due to factors such as gender, age, or regional background.
Which Google Cloud AI principle addresses this consideration of fairness, transparency, and accountability in AI systems?
Answer: B. Responsible AI
Responsible AI is Google Cloud's AI principle encompassing the fairness, transparency, and accountability required in this scenario, directly providing review criteria to ensure that the recruitment AI does not disadvantage specific groups based on gender, age, or region. Based on Google's AI principles, Responsible AI includes bias detection, fairness assurance, and management of the social impact of AI decision-making, supporting the recruitment platform in simultaneously fulfilling regulatory compliance and social responsibility.
Responsible AI is a comprehensive ethical principles framework applied when designing and operating AI systems in Google Cloud. Its core values are structured around 7 principles: Socially Beneficial, Avoid Unfair Bias, Safety, Accountability, Privacy Protection, Scientific Excellence, and Apply AI Principles consistently. Explainable AI is a sub-component of Responsible AI, a technical approach specialized in transparency for explaining the reasoning behind model decisions.
In the exam, distinguishing between Explainable AI and Responsible AI is essential. When keywords such as bias prevention, fairness, preventing discrimination against specific groups, and ethical AI principles appear, Responsible AI is the correct answer. When explaining why a model made a specific decision, transparency, and regulatory compliance are the focus, select Explainable AI. Vertex AI Model Monitoring is for detecting performance degradation and should not be confused with ethical principles.
A large call center operator wants to automatically transcribe customer calls into text for analysis.
Which Google Cloud service can be adopted MOST quickly without building a separate model?
Answer: D. Speech-to-Text API
A large call center needs to automatically convert call audio into text for analysis without building a separate model, requiring the FASTEST possible adoption. Speech-to-Text API provides Google's pre-trained speech recognition model as an API, allowing immediate integration without any ML expertise. A single REST API or gRPC call converts audio to text, making it faster to adopt than any other option.
Speech-to-Text API is a fully managed speech recognition service pre-trained using Google's deep learning technology. It supports over 80 languages and dialects, and handles both real-time streaming recognition and batch processing. A dedicated call center recognition model is also available, delivering high accuracy even with telephone-quality audio.
Vertex AI and AutoML are platforms for training custom models, requiring data collection, training, and deployment pipelines that take significant time to set up. When a scenario specifies 'no separate model development' or 'fastest adoption,' choose a pre-trained API such as Speech-to-Text API; when custom training is required, choose AutoML or Vertex AI.
A media company collects tens of thousands of customer reviews and news articles every day.
The company wants to automatically extract key information such as people's names, company names, and locations from this unstructured text data to use for content classification. The team has no machine learning expertise.
Which Google Cloud service BEST meets these requirements?
Answer: C. Cloud Natural Language API
Cloud Natural Language API can immediately extract named entities such as people's names, company names, and locations from unstructured text using only a REST API call without ML expertise, satisfying both core conditions of this scenario simultaneously. Google's pre-trained model on large-scale text data can be used with a single API key, requiring no model training or infrastructure setup.
The Entity Analysis feature of Cloud Natural Language API automatically identifies entity types such as PERSON, ORGANIZATION, and LOCATION from text and returns them with confidence scores. As a fully managed service, no scaling or infrastructure management is needed even when processing tens of thousands of requests, and sentiment analysis, syntax analysis, and content classification features are also available through the same API.
Option 1 (Vertex AI custom model) and Option 2 (AutoML Text Classification) each require ML expertise or custom training data, which does not meet the 'quick adoption, no expertise needed' condition. Option 4 (Dataflow text pipeline) is a data processing pipeline without built-in named entity recognition. When you see 'no ML expertise + immediate adoption + named entity recognition', choose Cloud Natural Language API.
A global airline operates a seat reservation system across multiple regions worldwide. Reservation data in Seoul, New York, and London must always remain in the same state, and the system must be able to process millions of transactions simultaneously.
Which Google Cloud database service BEST meets both strong consistency and global scalability requirements?
Answer: C. Cloud Spanner
This question tests whether you can identify a service that simultaneously offers global strong consistency and a horizontally scalable relational database. The requirement that reservation data in Seoul, New York, and London must always remain in the same state means strong consistency, not just replication. Cloud Spanner uses Google's TrueTime API to implement global clock synchronization, making it the only service that always returns the same latest data regardless of which region you read from.
Cloud Spanner is a fully managed distributed relational database on Google Cloud that guarantees a 99.999% availability SLA, standard SQL (ANSI 2011) support, and ACID transactions at a global scale. It can increase throughput without downtime by adding nodes via horizontal scaling and can process millions of concurrent transactions from a single instance. It is optimized for workloads where global consistency is business-critical, such as airline reservations, financial transactions, and game leaderboards.
On the GCP exam, when all three of 'global + strong consistency + relational' appear together, Cloud Spanner is the correct answer. The difference from Cloud SQL is that Cloud SQL operates only in a single region and does not guarantee global strong consistency. The difference from Firestore is that Firestore is a NoSQL database that does not support complex relational transactions.
An online shopping mall is experiencing repeated DDoS attacks and SQL injection attempts during large-scale sale events. The solution must minimize operational overhead.
Which Google Cloud service BEST protects against these web-based attacks?
Answer: A. Cloud Armor
Cloud Armor simultaneously meets both of this scenario's core requirements — DDoS attack mitigation and SQL injection/XSS blocking — without operational overhead. As a fully managed service combined with Google's global network infrastructure, it absorbs large-scale DDoS traffic and automatically blocks OWASP Top 10 web attacks through predefined WAF rule sets. It integrates natively with Cloud Load Balancing to inspect all inbound traffic in front of the HTTP(S) load balancer.
Cloud Armor's WAF capabilities include IP allow/deny lists, geographic filtering, rate limiting, and Adaptive Protection. Adaptive Protection uses machine learning to automatically detect abnormal traffic patterns and immediately propose mitigation rules when attacks occur. Based on Google's global-scale infrastructure, it can defend against volumetric DDoS attacks ranging from tens to hundreds of Gbps.
When 'DDoS' and 'SQL injection (or WAF)' appear together on the exam, Cloud Armor is the correct answer. VPC Service Controls (Option 2) is designed to prevent API data exfiltration, Cloud CDN (Option 3) is for performance optimization, and Cloud Load Balancing (Option 4) is for traffic distribution and does not provide security defense capabilities on its own.
The Google Cloud Digital Leader (CDL) exam consists of 50 questions with a 90-minute time limit.
The passing score for the CDL exam is approximately 70%.
The CDL sample set (20 questions with full explanations) is free. The full bank of 320+ questions is available with a CloudMasterIT subscription.
The CDL certification is valid for 3 years after passing. Recertification is required after that.