01. Which Vertex AI offering gives data scientists a managed notebook environment for exploring data and developing models?
a) Vertex AI Workbench
b) Vertex AI Agent Builder
c) Vertex AI Model Garden
d) Vertex AI Feature Store
02. What does a data catalog give an organization preparing for a gen AI project?
a) A consolidated copy of every dataset, gathered into one store and kept in step with its sources
b) An automatic assessment of whether each dataset is accurate
c) A searchable record of what data exists, where it sits and what it means
d) A set of access credentials for each system holding data
03. An organization holds a million support tickets but has budget to have only a few thousand of them labeled by hand.
Which approach lets the labeled portion guide learning across the whole collection?
a) Supervised learning, restricted to the few thousand labeled tickets and the categories they carry
b) Semi-supervised learning, using the small labeled set alongside the large unlabeled one
c) Reinforcement learning, which rewards the model for routing each new ticket correctly over time
d) Unsupervised learning, setting aside the labels that were produced
04. When a language model request is metered by token volume, which text is counted?
a) Only the text that the model generates and sends back
b) Only the text that the calling application itself sends in
c) Only the portion of the request that exceeds a free allowance
d) Both the text sent in and the text generated back
05. Before a large body of support transcripts is used for model work, an organization must find and mask any personal identifiers they contain.
Which Google Cloud capability is intended for this?
a) Identity and Access Management, which controls who may open the transcripts
b) Sensitive Data Protection, which discovers and de-identifies personal information
c) Cloud Logging, which records who accessed the transcripts and when
d) Security Command Center, which reports misconfigurations across the estate
06. Why is specialized hardware used to train large models rather than ordinary general-purpose processors?
a) It guarantees the trained model will produce reproducible answers
b) It stores the training dataset closer to where the model is served, and shortens the path between them
c) It allows the model to learn from fewer training examples
d) It performs the many parallel mathematical operations training requires far faster
07. Much of the information an enterprise holds sits in documents, recordings and images rather than in databases.
Why does generative AI change the value of that material?
a) It converts that material into database tables first, and discards whatever will not fit a column
b) It can interpret material that previously had to be read by a person to be useful
c) It reduces the storage that material occupies
d) It removes the need to govern access to that material
08. A prompt tells the model to summarize a report and adds that the summary must leave out any pricing figures.
What is this addition called?
a) A constraint on the output
b) A role assignment for the model to adopt
c) A worked example for the model to follow
d) A chain of reasoning steps to work through
09. Sensor readings arrive continuously and must be handled as they land, rather than collected and processed on a nightly schedule.
How is this data movement described?
a) Batch
b) Streaming
c) Federated
d) Archival
10. How do artificial intelligence, machine learning and generative AI relate to one another?
a) Machine learning sits inside generative AI, which in turn sits inside artificial intelligence
b) Artificial intelligence sits inside generative AI, which sits inside machine learning
c) Generative AI sits inside machine learning, which sits inside artificial intelligence
d) The three are separate fields of study that overlap only where they happen to share techniques
We have prepared Google Generative AI Leader (GCP-GAIL) certification sample questions to make you aware of actual exam properties. This sample question set provides you with information about the Google GCP-GAIL exam pattern, question formate, a difficulty level of questions and time required to answer each question. To get familiar with Google Cloud Platform - Generative AI Leader (GCP-GAIL) exam, we suggest you try our Sample Google GCP-GAIL Certification Practice Exam in simulated Google certification exam environment.
