A financial company is developing a generative AI application for loan approval decisions. The company needs the application output to be responsible and fair.
Which solution meets these requirements?
A. Review the training data to check for biases. Include data from all demographics in the training data.
B. Use a deep learning model with many hidden layers.
C. Keep the model’s decision-making process a secret to protect proprietary algorithms.
D. Continuously monitor the model’s performance on a static test dataset
Show Answer
Correct Answer: A
Explanation: Ensuring responsible and fair AI outputs starts with addressing bias at the source. Reviewing training data for bias and ensuring representation across all demographics reduces the risk of discriminatory loan decisions. The other options do not directly address fairness or responsibility: model complexity does not ensure fairness, secrecy reduces accountability, and monitoring on a static dataset does not prevent biased outcomes.
Question 26
An education company wants to build a private tutor application. The application will give users the ability to enter text or provide a picture of a question. The application will respond with a written answer and an explanation of the written answer.
Which model type meets these requirements?
A. Computer vision model
B. Multimodal LLM
C. Diffusion model
D. Text-to-speech model
Show Answer
Correct Answer: B
Explanation: The application must accept both text input and image input (a picture of a question) and generate a written answer with an explanation. This requires understanding and reasoning across multiple input modalities and producing text output, which is exactly the capability of a multimodal large language model. Computer vision alone cannot generate explanations, diffusion models generate images, and text-to-speech converts text to audio.
Question 27
A company needs to collect a large dataset to train an AI assistant in a specific content area.
Which dataset will meet this requirement?
A. Diverse conversations that use relevant terminology
B. Time series data of general purpose historical sales
C. Sentiment analysis of news articles
D. Unique product IDs and corresponding user IDs
Show Answer
Correct Answer: A
Explanation: Training an AI assistant for a specific content area requires large volumes of conversational data that reflect how users ask questions and how responses should be formed, using domain-specific terminology. Diverse, relevant conversations directly support learning language patterns, context, and subject-matter usage, unlike time series, sentiment labels, or ID mappings.
Question 27
A company trains image and text generation models on Amazon SageMaker AI. The company releases the models by using Amazon Bedrock. The company must retain a tamper-proof, queryable record of every API call from SageMaker AI, Amazon Bedrock, and AWS Identity and Access Management (IAM).
Which AWS service will meet these requirements?
A. AWS Trusted Advisor
B. Amazon Macie
C. AWS CloudTrail Lake
D. Amazon Inspector
Show Answer
Correct Answer: C
Explanation: AWS CloudTrail Lake provides a tamper‑resistant, immutable event store and supports SQL‑based querying of API activity. It records and retains API calls from SageMaker AI, Amazon Bedrock, and IAM, meeting the requirement for a tamper‑proof, queryable audit record. The other services do not provide comprehensive API call logging and querying.
Question 28
A news agency publishes articles in English. The agency wants to make articles available in other languages.
Which solution meets these requirements?
A. Add Amazon Transcribe to the company’s website.
B. Use the Amazon Translate real-time translation feature.
C. Add Amazon Personalize to the company’s website.
D. Use the Amazon Textract real-time document processing feature.
Show Answer
Correct Answer: B
Explanation: The requirement is to translate existing English text into other languages. Amazon Translate is designed to perform automatic language translation of text in real time. The other services address different problems: Amazon Transcribe converts speech to text, Amazon Personalize provides recommendations, and Amazon Textract extracts text from documents.
Question 28
HOTSPOT
-
A company wants to customize a foundation model (FM). The company wants to understand the customization methods and data types that are available.
Select the correct customization method from the following list for each description. Select each customization method one time.
Show Answer
Correct Answer: Provide labeled data to improve performance on specific tasks:
Fine-tuning
Provide unlabeled data for a specific domain:
Continued pre-training
Transfer knowledge from a larger model to a smaller model:
Distillation
Explanation: Fine-tuning uses labeled task-specific data to adapt model behavior. Continued pre-training uses large amounts of unlabeled domain data to adapt the model’s representations. Distillation compresses knowledge from a larger teacher model into a smaller student model.
Question 29
A company has guidelines for data storage and deletion.
Which data governance strategy does this describe?
A. Data de-identification
B. Data quality standards
C. Data retention
D. Log storage
Show Answer
Correct Answer: C
Explanation: Guidelines that define how long data is stored and when it should be deleted are part of a data retention strategy. Data retention policies govern storage duration and disposal, whereas de-identification, data quality standards, and log storage address different aspects of data governance.
Question 29
A company wants to increase employee productivity by using a generative AI solution to write code to test software applications.
Which solution will meet these requirements with the LEAST operational effort?
A. Amazon Q Business
B. Amazon Bedrock Agents
C. Amazon Q Developer
D. Amazon SageMaker Clarify
Show Answer
Correct Answer: C
Explanation: Amazon Q Developer is purpose-built to help developers write, review, and test code using generative AI with minimal setup. It integrates directly into IDEs and AWS services, requiring far less operational effort than building custom agents with Amazon Bedrock, deploying enterprise search with Amazon Q Business, or using SageMaker Clarify, which is for ML bias and explainability rather than code generation.
Question 30
An AI practitioner has trained a model on a training dataset. The model performs well on the training data. However, the model does not perform well on evaluation data.
What is the MOST likely cause of this issue?
A. The model is underfit.
B. The model requires prompt engineering.
C. The model is biased.
D. The model is overfit.
Show Answer
Correct Answer: D
Explanation: When a model performs very well on the training data but poorly on evaluation (validation/test) data, it indicates the model has learned patterns specific to the training set rather than generalizable features. This is the classic symptom of overfitting. An underfit model would perform poorly on both training and evaluation data.
Question 30
A company wants to use large language models (LLMs) to create a chatbot. The chatbot will assist customers with product inquiries, order tracking, and returns. The chatbot must be able to process text inputs and image inputs to generate responses.
Which AWS service meets these requirements?
A. Amazon Bedrock
B. Amazon Comprehend
C. Amazon Q
D. Amazon Rekognition
Show Answer
Correct Answer: A
Explanation: Amazon Bedrock provides access to foundation models that can handle both text and image inputs and can be used to build multimodal chatbots powered by large language models. The other services focus on narrower tasks (Comprehend for NLP analysis, Q for enterprise assistance, Rekognition for image analysis only) and do not meet the full requirement.
$19
Get all 397 questions with detailed answers and explanations
Instant download HTML + PDF delivered the moment payment clears.
Secure Stripe checkout we never see or store your card details.
7-day refund if files are defective see our refund policy.
