Amazon Practice Questions, Discussions & Exam Topics by our Authors
Which option is a use case for generative AI models?
Let's break down each option based on the use case for generative AI models, while considering services, effort, time, cost, and other key factors.
Option A: Improving network security by using intrusion detection systems
Intrusion detection systems (IDS) typically rely on pattern recognition, rule-based systems, and machine learning models to detect potential threats. While machine learning models can be used here, generative AI is not typically the best fit. The effort, time, and cost of implementing generative AI for network security wouldn't yield substantial benefits over existing non-generative models. The service involved here is more related to anomaly detection, which isn’t a primary use case for generative AI.
Rejection reason: Generative AI doesn't naturally apply to real-time network security monitoring and anomaly detection, as its strength lies in content generation and data augmentation.
Option B: Creating photorealistic images from text descriptions for digital marketing
Generative AI, especially models like GPT, DALL·E, and other text-to-image generators, excel at creating photorealistic images from textual descriptions. This can significantly reduce the time, effort, and cost associated with content creation for digital marketing. For example, it can help marketers quickly generate visual assets without needing photographers or designers, thus saving time and resources.
Selected reason: Generative AI is well-suited for this use case. It saves time and costs by automating the image creation process and enhances the creativity of digital marketing campaigns.
Option C: Enhancing dat...
Author: Ethan · Last updated May 7, 2026
A company wants to build a generative AI application by using Amazon Bedrock and needs to choose a foundation model (FM). The company wants to know how much information can ...
To determine how much information can fit into one prompt, the company needs to consider factors that directly affect the capacity of a model to process the input data. Let’s evaluate each option with respect to the question and the key considerations such as effort, time, cost, and other relevant factors:
A) Temperature
The temperature setting in generative AI controls the randomness of the output, influencing the creativity of responses. A higher temperature results in more diverse outputs, while a lower temperature produces more deterministic responses. However, temperature does not affect how much information can be fed into the model in a single prompt. It only impacts the behavior of the model’s output, not its input capacity.
Pros: Affects output diversity.
Cons: Not relevant to the amount of information that can fit into one prompt.
B) Context window
The context window refers to the number of tokens the model can process in a single prompt. This is the critical factor when determining how much information can be included in one prompt. A larger context window allows for a more extensive input (more tokens) to be fed into the model, whereas a smaller context window limits the amount of information that can be processed at once. In the context of Amazon Bedrock and building generative AI applications, this is the most relevant consideration when deciding how much information can fit into one prompt.
Pros: Directly determines how much input (text) can be processed.
Cons: Larger context windows may require more compute power, potentially increasing cost and resource usage.
C) Batch size
Batch size refers to the number of input requests pro...
Author: Isabella · Last updated May 7, 2026
A company wants to make a chatbot to help customers. The chatbot will help solve technical problems without human intervention.The company chose a foundation model (FM) for the chatbot. The chatbot needs t...
Let's analyze the options based on the requirements: creating a chatbot that adheres to the company tone while solving technical problems without human intervention using a foundation model (FM).
A) Set a low limit on the number of tokens the FM can produce:
Limiting the number of tokens produced by the FM could make responses more concise, but it won't directly ensure that the responses adhere to the company’s tone or solve technical problems effectively. The length of the response is important, but it doesn't guarantee that the content or tone will match the company's needs. The tone and accuracy are more about how the FM is instructed or fine-tuned to produce outputs, not just the token count. This approach doesn’t address the core requirement of ensuring a consistent tone.
B) Use batch inferencing to process detailed responses:
Batch inferencing is useful for processing large amounts of data in parallel and can save costs and time in some cases, but it's not specifically designed to handle real-time, interactive chatbot responses. Chatbots require quick, real-time interactions with users, not batch processing. This solution would be inefficient for providing timely responses and does not specifically address the requirement of ensuring a consistent tone in individual interactions.
C) Experiment and refine the prompt until the FM produces the desired responses:
Experimenting with and refining the prompt is the most relevant and effective approach. By adjusting the prompt, you ca...
Author: Olivia Johnson · Last updated May 7, 2026
A company wants to use a large language model (LLM) on Amazon Bedrock for sentiment analysis. The company wants to classify the sentiment of text passages as positive...
Let's break down the options in terms of efficiency, cost, time, and accuracy for sentiment analysis using Amazon Bedrock with an LLM.
Option A: Provide examples of text passages with corresponding positive or negative labels in the prompt followed by the new text passage to be classified.
- Efficiency: This is a highly effective strategy for LLM-based sentiment analysis. Providing labeled examples directly allows the model to learn from context and apply those learned patterns to classify new text accurately.
- Cost: This is cost-effective, especially since Amazon Bedrock services would benefit from providing examples upfront. The cost is generally low in terms of computing power, as the model only needs to classify based on patterns it has already been exposed to in the prompt.
- Time: This approach saves time compared to more complex strategies. With relevant labeled examples, the model can quickly classify new text.
- Key Factor: It focuses directly on the task and reduces ambiguity, improving the model's accuracy and speed in classification. The only extra "effort" required is choosing good, representative examples to use in the prompt.
Option B: Provide a detailed explanation of sentiment analysis and how LLMs work in the prompt.
- Efficiency: This would increase the prompt size unnecessarily. LLMs can perform sentiment analysis without needing an exhaustive explanation. The explanation is more about educating the model, which is unnecessary for a basic task like sentiment classification.
- Cost: More words in the prompt result in higher compute cost since more data needs to be processed.
- Time: Time to process this option will be longer because the explanation requires the model to parse and potentially understand irrelevant details.
- Key Factor: This option doesn’t add value for the specific task of sentiment analysis. It is better suited for tasks where th...
Author: SolarFalcon11 · Last updated May 7, 2026
A security company is using Amazon Bedrock to run foundation models (FMs). The company wants to ensure that only authorized users invoke the models. The company needs to identify any unauthorized access attempts to set appropriate AWS Identity and Access Management (IAM) policies and roles for future ite...
To determine the best AWS service for identifying unauthorized access attempts to Amazon Bedrock, we need to evaluate the capabilities of each service based on the use case, which involves tracking and identifying unauthorized users interacting with Amazon Bedrock. Let’s analyze each option:
A) AWS Audit Manager
- Operational Overhead: AWS Audit Manager helps automate audit preparations by collecting evidence and generating reports. It is primarily used for regulatory compliance, such as ensuring adherence to standards like SOC 2 or GDPR.
- Suitability: Although AWS Audit Manager provides auditing capabilities, it does not specialize in tracking real-time access or identifying unauthorized access attempts to specific AWS services.
- Conclusion: Not ideal. It's more focused on compliance and auditing than monitoring unauthorized access attempts in real time.
B) AWS CloudTrail
- Operational Overhead: AWS CloudTrail records API calls made on your AWS account, including those for Amazon Bedrock. This service allows you to track who accessed what, when, and from where, enabling you to identify unauthorized access attempts or actions.
- Suitability: CloudTrail is specifically designed for tracking API activity across AWS services. It logs every request to your AWS resources, making it ideal for detecting unauthorized access, including to Amazon Bedrock.
- Cost and Time: CloudTrail has minimal overhead in terms of configuration. It can be set up quickly, and costs are typically based on the number of events logged. The service is highly efficient for tracking unauthorized access in real time.
- Conclusion: Best option. It provides the necessary real-time visibility into access attempts, is easy to implement, and aligns with the company's needs for detecting unauthorized users.
C) Amazon Fraud Dete...
Author: Ella · Last updated May 7, 2026
A company has developed an ML model for image classification. The company wants to deploy the model to production so that a web application can use the model.The company needs to implement a solution to host the model and serve predicti...
To determine the best solution, let's carefully analyze each option based on the given requirements: no management of infrastructure, serving predictions for a web application, time, effort, cost, and relevant services.
Option A: Amazon SageMaker Serverless Inference
- Description: Amazon SageMaker Serverless Inference allows you to deploy machine learning models without managing infrastructure. It automatically scales to handle inference requests, and you only pay for the compute time used during inference.
- Key factors:
- Infrastructure management: No infrastructure management required; SageMaker handles scaling and provisioning.
- Effort & time: Minimal effort required to set up; it’s designed for simplicity and ease of use for deploying ML models.
- Cost: Pay-as-you-go model, which is cost-effective as you only pay for actual usage.
- Scalability: Serverless scaling means the service can handle varying loads without manual intervention.
- Why it's suitable: It directly meets the requirement of serving ML predictions in production without infrastructure management. It's designed for this exact use case: hosting models and serving predictions at scale.
Option B: Amazon CloudFront
- Description: Amazon CloudFront is a content delivery network (CDN) service designed to deliver static and dynamic content (like web pages, images, etc.) with low latency.
- Key factors:
- Infrastructure management: No machine learning model deployment support.
- Effort & time: You would have to manually deploy and manage the model in a different service (like EC2) and then use CloudFront to cache and deliver static assets or API responses.
- Cost: CloudFront charges based on data transfer, but it’s not suitable for serving machine learning models directly.
- Why it's not suitable: CloudFront is not a solution for hosting or serving ML models; it's a CDN. It can't serve as the model-serving platform itself. While useful for web content distribution, it doesn’t fit the use case of serving machine learning predictions directly.
Option C:...
Author: Liam123 · Last updated May 7, 2026
An AI company periodically evaluates its systems and processes with the help of independent software vendors (ISVs). The company needs to receive email message notifications when an ISV's compliance r...
To meet the requirement of receiving email notifications when an ISV's compliance reports become available, let's evaluate the options based on services, effort, time, cost, and other key factors.
A) AWS Audit Manager
AWS Audit Manager helps you automate the collection of evidence related to compliance audits. It tracks and assesses your organization's compliance with regulations. While this service is great for auditing purposes, it’s more focused on gathering evidence for internal assessments rather than receiving notifications about external compliance reports, such as from an ISV. Therefore, it would not be the most suitable option for receiving notifications about ISV reports.
B) AWS Artifact
AWS Artifact is the most relevant service in this case. It is a service that provides on-demand access to AWS's compliance reports, such as certifications and audits. It also allows you to view and download compliance reports from AWS and other third-party vendors. However, AWS Artifact doesn't automatically send email notifications when new reports are available. If notifications are needed, the company might have to manually check for updates within the Artifact portal, or potentially set up custom workflows for notifications. Despite this limitation, AWS Artifact is the closest fit for retrieving compliance reports.
C) AWS Trusted Advisor
AWS Trusted Advisor is a service that provid...
Author: Krishna · Last updated May 7, 2026
A company wants to use a large language model (LLM) to develop a conversational agent. The company needs to prevent the LLM from being manipulated with common prompt engineering techniques to perform un...
To address the company’s goal of preventing the LLM from being manipulated by common prompt engineering techniques, we need to consider factors like effectiveness, cost, time, and other technical considerations. Let’s break down each option:
A) Create a prompt template that teaches the LLM to detect attack patterns.
This approach aims to help the LLM recognize malicious or manipulative prompt patterns, making it resistant to prompt engineering attacks. The company would need to invest time and effort in developing and fine-tuning these templates. However, the effectiveness of this approach depends on how well the template captures all attack vectors, which can be difficult given the constantly evolving nature of prompt engineering techniques. This also involves ongoing maintenance and updates as new attack strategies emerge, which can incur additional costs.
Pros: This could be highly effective if implemented well and continuously updated.
Cons: Requires significant ongoing effort and time for maintenance and optimization.
B) Increase the temperature parameter on invocation requests to the LLM.
The temperature parameter controls the randomness of the model’s outputs. Higher temperature values lead to more diverse outputs, but they can also make the LLM more unpredictable and potentially less safe. If the goal is to prevent the model from being manipulated or providing undesired outputs, increasing the temperature could backfire and make the model more susceptible to undesired responses. This option does not address the core issue of prompt manipulation but could worsen control over the output.
Pros: Might make the model less repetitive.
Cons: It increases randomness and could make the model more difficult to control, which is not desirable when trying to prevent manipulation.
C) Avoid using LLMs that are not listed in Amazon SageMaker.
Amazon Sage...
Author: Ava · Last updated May 7, 2026
A company is using the Generative AI Security Scoping Matrix to assess security responsibilities for its solutions. The company has identified four different solution scopes based on the matrix.Wh...
When assessing security responsibilities in the context of the Generative AI Security Scoping Matrix, the option where the company takes the most ownership of security responsibilities is the one where it has the most control over the entire process, from model development to deployment and ongoing operation. Let’s break down each option:
A) Using a third-party enterprise application that has embedded generative AI features:
- In this scenario, the company relies on a third-party to provide the enterprise application, which includes generative AI features. The responsibility for security in this case is largely on the third-party provider, as they manage the application and the embedded AI features. The company’s ownership of security responsibilities is limited to how the application is configured and used, not the underlying AI models.
B) Building an application by using an existing third-party generative AI foundation model (FM):
- This option still involves using a third-party foundation model. While the company has more control over the application layer and how it interacts with the model, the core generative AI model itself remains external. The company’s security responsibilities are increased in comparison to option A, but not to the extent of having full control over the entire solution.
C) Refining an existing third-party generative AI foundation model (FM) by fine-tuning the model by using data specific to the business:
- Fine-tuning an existing third-party model...
Author: Henry · Last updated May 7, 2026
An AI practitioner has a database of animal photos. The AI practitioner wants to automatically identify and categorize the animals in the photos with...
To determine the most suitable strategy for automatically identifying and categorizing animals in a database of photos, let’s evaluate the options based on the given requirements: automatically identify and categorize the animals without manual human effort.
Option A: Object detection
- Description: Object detection is a computer vision task that involves detecting and localizing objects in images or videos. It also classifies the objects it detects (e.g., animals in this case).
- Key factors:
- Relevance: Object detection is directly suited for identifying and categorizing objects (like animals) in images. This task aligns perfectly with the goal of the practitioner: identifying and categorizing animals in photos.
- Effort & time: Implementing an object detection model requires initial training on labeled data (if a pre-trained model is not used). After that, inference can be run automatically without manual intervention.
- Cost: Training an object detection model may incur costs depending on the size and complexity of the data, but pre-trained models (like those in Amazon SageMaker or TensorFlow) can reduce both cost and time significantly.
- Why it's suitable: This strategy allows the AI practitioner to automatically identify and categorize animals, making it the best fit for the given problem. Pre-trained object detection models can handle a wide range of categories, including animals, and can be fine-tuned for better accuracy.
Option B: Anomaly detection
- Description: Anomaly detection is a technique used to identify unusual or unexpected patterns in data. In the context of images, it’s often used to detect outliers or abnormalities in the data.
- Key factors:
- Relevance: Anomaly detection is not suited for identifying and categorizing objects. It...
Author: Carlos Garcia · Last updated May 7, 2026
A company wants to create an application by using Amazon Bedrock. The company has a limited budget and prefers flexibility without long-term commitm...
When considering the best pricing model for the company's requirements on Amazon Bedrock, the company has a limited budget and prefers flexibility without long-term commitment. Let’s evaluate each option:
Option A: On-Demand
- Reasoning: On-Demand pricing allows the company to pay for the resources they use without committing to long-term contracts or upfront costs. This model offers flexibility because the company only pays for the capacity they consume, which is ideal for businesses with variable workloads or a limited budget. They can scale up or down as needed, paying for the exact amount of resources used, without any long-term financial commitments.
- Effort: Low. The company can easily manage their usage and costs on a pay-as-you-go basis.
- Time: Low. Setting up on-demand services is quick and flexible.
- Cost: Variable. Costs depend on usage, but it is cost-effective for companies with limited budgets and fluctuating needs.
- Why selected: On-Demand pricing perfectly fits the company's need for flexibility without long-term commitments, and it allows them to manage costs based on actual usage.
Option B: Model customization
- Reasoning: Model customization allows for tailoring pre-built models to better suit specific business needs. While this can enhance the model’s performance, it typically involves higher costs due to the need for specialized expertise and resources for customization. This might not align with the company’s goal of keeping costs low and flexible.
- Effort: High. Customizing models requires specialized knowledge and time.
- Time: Medium to high. Customization can be a lengthy process, especially if significant adjustments are needed.
- Cost: High. There are additional costs associated with model customization, making it less suitable for a limited budget.
- Why rejected: This model is better for companies with more resources and a need for highly specialized models. It doesn’t fit with the company's preference for a flexible, bud...
Author: Zara · Last updated May 7, 2026
Which AWS service or feature can help an AI development team quickly deploy and consume a foundation...
Let's analyze each option based on the context of quickly deploying and consuming a foundation model (FM) within the team's Virtual Private Cloud (VPC). We'll consider factors like services, effort, time, cost, and other key factors in reasoning.
Option A: Amazon Personalize
Amazon Personalize is an AWS service designed specifically for building recommendation systems, such as personalized product recommendations or content suggestions. While it uses machine learning models, it is not a service intended for quickly deploying and consuming a foundation model (FM). It focuses on custom models for recommendation, not general-purpose foundation models.
Rejection reason: Amazon Personalize is not meant for quickly deploying foundation models or using pre-trained models for general AI tasks, so it's not suitable for the given requirement.
Option B: Amazon SageMaker JumpStart
Amazon SageMaker JumpStart is a feature of SageMaker that helps developers quickly start and experiment with machine learning models, including pre-trained models and foundation models. JumpStart provides a curated set of models, including FMs, and makes it easier to deploy them within your environment, including a VPC. This service significantly reduces the effort, time, and cost needed to get a foundation model up and running in a team’s environment, allowing quick experimentation and deployment.
Selected reason: SageMaker JumpStart is specifically designed to help teams quickly deploy and consume foundation models, fitting the exact use case described. It also allows integration with VPCs for secure, scalable deployment.
Option C: PartyRock, an Amazon Bedrock Playground
Part...
Author: Evelyn · Last updated May 7, 2026
How can companies use large language models (LLMs) securely on Amazon Bedrock?
When companies look to securely use large language models (LLMs) on Amazon Bedrock, the approach involves evaluating how to ensure both security and effectiveness in terms of cost, effort, time, and key factors such as governance, explainability, and access control. Let's break down each option:
Option A: Design clear and specific prompts. Configure AWS Identity and Access Management (IAM) roles and policies by using least privilege access.
- Reasoning: This option focuses on the foundational security and governance aspect of using LLMs. Configuring IAM roles with least privilege access ensures that users and systems only have access to the resources they need, minimizing the risk of unauthorized access. Designing specific prompts is essential to guiding the models effectively, improving their accuracy, and reducing the potential for misuse.
- Effort: Medium. Designing secure IAM roles takes time, and prompt engineering requires understanding the task and data.
- Time: Medium. Setting up IAM roles and designing prompts can take time but ensures long-term security.
- Cost: Low to medium. There may be some overhead in managing IAM configurations, but this is mostly a one-time setup cost.
- Why rejected: This is not rejected—this approach is critical for ensuring security and effectiveness.
Option B: Enable AWS Audit Manager for automatic model evaluation jobs.
- Reasoning: AWS Audit Manager is useful for auditing and tracking compliance, especially in regulated industries. While it helps with tracking and governance, it doesn't specifically address the security aspects of model deployment or the direct use of LLMs on Bedrock.
- Effort: Low. AWS Audit Manager automates much of the auditing process.
- Time: Medium. Initial configuration of AWS Audit Manager is required.
- Cost: Medium to high, depending on the number of resources being audited.
- Why rejected: This is valuable for comp...
Author: Aarav · Last updated May 7, 2026
A company has terabytes of data in a database that the company can use for business analysis. The company wants to build an AI-based application that can build a SQL query from input text that employees provide. Th...
To address the company's requirement of building an AI-based application that can generate SQL queries from input text, let’s analyze each option:
Option A: Generative pre-trained transformers (GPT)
- Reasoning: GPT, such as the model behind ChatGPT, is highly suitable for this use case because it excels at natural language processing tasks. GPT can understand and generate text in human language, which makes it capable of translating input text into SQL queries. GPT has been shown to be effective in tasks like query generation and can be easily fine-tuned for specific tasks, including understanding business requirements from minimal input and translating them into SQL syntax.
- Effort: Medium. Integrating GPT into the application may require some development work, especially to fine-tune the model for the specific use case.
- Time: Medium to high. Setting up the model and integrating it with the company's existing data infrastructure (database access, query execution) will take time.
- Cost: Medium. Using a GPT-based model like OpenAI’s GPT might incur usage costs based on how often employees use the application, but these costs are scalable.
- Why selected: GPT is the best option here because of its strong natural language processing capabilities, which are perfect for generating SQL queries from text input. Additionally, it is flexible and can be customized with fine-tuning if necessary to suit the company's specific needs.
Option B: Residual neural network
- Reasoning: A residual neural network (ResNet) is typically used for tasks such as image recognition or other applications that require deep convolutional networks. It is not suited for natural language processing or query generation tasks, which is the core requirement in this scenario.
- Effort: High. Building and training a ResNet for this purpose would require substantial effort and wouldn’t be effective for the text-to-SQL task.
- Time: High. Even with modifications, it would take a considerable amount of time to adapt a ResNet for language-related tasks.
- Cost: High. Given the unnecessary complexity ...
Author: Olivia · Last updated May 7, 2026
A company built a deep learning model for object detection and deployed the model to production.Which AI process occurs whe...
Let's analyze each option based on the context of a deep learning model deployed for object detection in production. We're considering the services, effort, time, cost, and the exact process described in the question.
Option A: Training
Training is the process where the model learns from labeled data. During training, the model adjusts its parameters to minimize the error on a given task (e.g., object detection). However, this happens before the model is deployed to production. Once the model is deployed, it no longer undergoes training in the traditional sense (unless you are retraining the model with new data).
Rejection reason: The process described in the question refers to the model analyzing a new image to identify objects, which happens after training, during actual usage. Thus, it isn't about training.
Option B: Inference
Inference refers to the process where a trained model is used to make predictions on new, unseen data. In this case, the model is analyzing a new image and identifying objects, which is exactly what inference is. It is a runtime process where the trained model is deployed in production to perform its task of detecting objects in real-time or on batch data.
Selected reason: This directly aligns with the task described in the question. Inference happens when the model is deployed and is used to make predictions on new data, such as ident...
Author: Julian · Last updated May 7, 2026
An AI practitioner is building a model to generate images of humans in various professions. The AI practitioner discovered that the input data is biased and that specific attributes affect the i...
The AI practitioner is building a model to generate images of humans in various professions and has discovered that the input data is biased, with certain attributes influencing the generated images. This bias in the data is causing the model to produce biased results. Let’s evaluate each option to determine which technique will effectively solve the problem of bias in the model.
Option A: Data augmentation for imbalanced classes
- Explanation: Data augmentation involves artificially increasing the size of the training dataset by applying transformations to the existing data, such as rotating, flipping, or changing the lighting of images. For imbalanced classes, data augmentation can help balance the number of examples for different classes.
- Relevance: This option can help address bias if the bias is due to imbalanced representation of certain attributes (e.g., professions, gender, or ethnicity). By augmenting underrepresented categories, you can balance the dataset, making the model more likely to generate diverse and fair representations.
- Use case: This option is ideal when there is a significant imbalance in the data for certain classes (e.g., some professions are overrepresented in the dataset), as it would help mitigate biased predictions caused by the imbalance.
Option B: Model monitoring for class distribution
- Explanation: Model monitoring involves tracking the performance of the model over time, focusing on aspects like class distribution, accuracy, and fairness. Monitoring class distribution can highlight when the model's predictions are biased towards certain attributes or classes.
- Relevance: While monitoring class distribution can help detect bias in predictions, it doesn’t directly solve the issue. It is more of a diagnostic tool rather than an active technique to address and correct the bias in the model.
- Use case: This option is useful for detecting bias over time but doesn't actively solve the problem of biased data during the model training or generati...
Author: GlowingTiger · Last updated May 7, 2026
A company is implementing the Amazon Titan foundation model (FM) by using Amazon Bedrock. The company needs to supplement the model by using relevant data from the c...
To meet the company’s requirement of supplementing the Amazon Titan foundation model (FM) with relevant data from the company’s private data sources, we need a solution that allows the integration of private data with the model to enhance its capabilities. Let's evaluate each option:
A) Use a different FM
Switching to a different foundation model might not be necessary, as Amazon Titan can potentially meet the company's needs. The core issue here is integrating private data sources with the existing model, not changing the model itself. Using a different FM could involve significant effort and additional costs to retrain or adapt the model to the company's needs, making it less efficient than addressing the data integration requirement directly.
Pros: Might be relevant if Titan doesn’t fit the company's needs (e.g., specific task suitability).
Cons: Not directly related to the integration of private data and would involve unnecessary overhead.
B) Choose a lower temperature value
The temperature parameter controls the randomness of the model’s output. A lower temperature makes the responses more deterministic and controlled, but it does not influence the ability of the model to incorporate or learn from external data sources. Lowering the temperature may improve response consistency, but it doesn’t address the core requirement of supplementing the Titan model with private data for better contextual responses.
Pros: Helps make responses more consistent and predictable.
Cons: Does not enable the model to access or integrate private data sources.
C) Create an Amazon Bedrock knowledge base
Creating a knowledge base within Amazon Bedrock is a highly effective solution for supp...
Author: NightmareDragon2025 · Last updated May 7, 2026
A medical company is customizing a foundation model (FM) for diagnostic purposes. The company needs the model to be transparent and explainable to meet re...
To meet the regulatory requirement of transparency and explainability for the foundation model in a medical company, we must focus on providing interpretability of model decisions, which is crucial for compliance in diagnostic settings.
Let's go through each option:
- Option A: Configure the security and compliance by using Amazon Inspector.
- Purpose: Amazon Inspector is a service that helps assess the security and compliance posture of applications deployed on AWS. However, it focuses on vulnerabilities and compliance related to security, not on model explainability or interpretability.
- Reason for rejection: This option doesn't address the need for model explainability or transparency; it is more about ensuring security and compliance in the environment.
- Option B: Generate simple metrics, reports, and examples by using Amazon SageMaker Clarify.
- Purpose: Amazon SageMaker Clarify is specifically designed for improving the transparency and explainability of machine learning models. It helps to detect bias, generate explainable reports, and visualize how a model makes decisions. This solution is tailored for use in regulated industries, such as healthcare, where model transparency is crucial.
- Reason for selection: This directly addresses the regulatory requirement of transparency and explainability. It allows the company to understand how the model is making predictions and ensures that the model is fair and unbiased, which is critical in medical diagnostics.
- Additional factors: The solution is integrated with AWS services and can be used to monitor and explain machine learning models in ...
Author: Liam · Last updated May 7, 2026
A company wants to deploy a conversational chatbot to answer customer questions. The chatbot is based on a fine-tuned Amazon SageMaker JumpStart model. The application must comply with multiple regul...
In this case, the company is deploying a conversational chatbot based on a fine-tuned Amazon SageMaker JumpStart model, and it must comply with multiple regulatory frameworks. We need to identify which capabilities help ensure regulatory compliance. Let's analyze the options based on compliance, services, effort, time, and other factors.
A) Auto scaling inference endpoints
Auto-scaling inference endpoints in Amazon SageMaker provide the ability to automatically adjust the compute capacity for the chatbot based on the volume of requests. While auto-scaling is crucial for performance, cost efficiency, and handling traffic spikes, it does not directly help the company with compliance to regulatory frameworks. Compliance usually requires controlling and securing data, ensuring privacy, and meeting regulatory standards, which is beyond the scope of auto-scaling alone. Hence, this option is not directly relevant to meeting compliance requirements.
B) Threat detection
Threat detection is an essential capability to ensure that the chatbot’s system is protected from external attacks, data breaches, or unauthorized access. In the context of regulatory compliance, particularly in industries like finance, healthcare, or data-sensitive environments, it is crucial to detect and mitigate threats to protect customer data and privacy. Amazon services such as Amazon GuardDuty or AWS Security Hub can be used to monitor and detect threats in the application infrastructure, helping demonstrate compliance with regulatory frameworks that require security and data protection. This capability directly supports compliance efforts.
C) Data protection
Data protection is a fundamental aspect of regulatory compliance. Many regulatory frameworks (like GDPR, HIPAA, etc.) require that sensitive customer data be protected during storage, processing, and transmission. Amazon SageMaker offers features like encryption at rest and in transit, access control, and data masking, wh...
Author: Aria · Last updated May 7, 2026
A company is training a foundation model (FM). The company wants to increase the accuracy of the model up to a specific accept...
To determine the best solution for increasing the accuracy of a foundation model (FM) during training, we need to evaluate each option based on the goal of achieving a specific accuracy level. The process of improving a model’s accuracy typically involves adjusting various hyperparameters, model architectures, or training techniques. Let's examine each option:
Option A: Decrease the batch size
- Description: Batch size refers to the number of training samples used in one iteration before the model’s weights are updated. Decreasing the batch size means the model will update weights more frequently, potentially leading to more fine-grained updates.
- Key factors:
- Training dynamics: Smaller batch sizes can improve generalization because they introduce more noise into the gradient descent process, which can prevent overfitting. However, they can also make the training process slower and less stable.
- Effort & time: It may require more iterations for convergence, thus increasing training time. It may also require tuning other parameters like learning rate.
- Cost: Decreasing the batch size could increase training time and computational costs, as more iterations are required to complete the same amount of training.
- Why it might not be the best option: While smaller batch sizes can help with generalization, they don’t always directly improve accuracy and may introduce instability in training. Decreasing the batch size is often more beneficial in preventing overfitting than increasing accuracy.
Option B: Increase the epochs
- Description: Epochs refer to the number of times the entire dataset is passed through the model during training. Increasing the number of epochs can allow the model to learn more from the data and potentially improve accuracy.
- Key factors:
- Training dynamics: More epochs generally give the model more opportunities to learn, which can lead to better accuracy, assuming the model is not overfitting. However, after a certain point, more epochs can lead to diminishing returns or even overfitting.
- Effort & time: Increasing epochs will increase the training time since the model is exposed to the data more times.
- Cost: More training epochs result in higher computational costs and longer time to reach the desired accuracy level.
- Why it might be suitable: If the model is underfitting (i.e., not learning enough from the d...
Author: ShadowWolf101 · Last updated May 7, 2026
A company is building a large language model (LLM) question answering chatbot. The company wants to decrease the number of actions call center employees need to take to respond to customer questions.Whi...
The company is building a large language model (LLM) question answering chatbot and wants to decrease the number of actions call center employees need to take in responding to customer questions. The goal is to evaluate how effectively the chatbot can reduce employee intervention, leading to efficiency gains.
Let's assess each option:
Option A: Website engagement rate
- Purpose: Website engagement rate typically measures how much time users spend on a website or how often they interact with specific features (e.g., page views, clicks). While website engagement could be an indicator of user interest, it is not directly related to evaluating the effectiveness of a chatbot in reducing the workload of call center employees.
- Reason for rejection: This metric does not focus on the core business objective of reducing call center actions or improving the efficiency of customer service. It is more about user behavior on a website, which is not aligned with the goal of assessing call center impact.
Option B: Average call duration
- Purpose: Average call duration measures the average length of customer calls. This metric directly correlates to the efficiency of call center interactions. If the LLM chatbot successfully handles customer queries and reduces the need for employees to take further actions, we would expect the average call duration to decrease.
- Reason for selection: This is the most relevant business objective for evaluating the chatbot's effect. A decrease in average call duration indicates that the chatbot is effectively answering customer queries, thus reducing the actions required by call center employees. The shorter the call durati...
Author: StarryEagle42 · Last updated May 7, 2026
A company is developing a new model to predict the prices of specific items. The model performed well on the training dataset. When the company deployed the model to production, the model's perf...
To mitigate the problem of the model’s performance decrease in production, the company needs to focus on ensuring that the model generalizes well to real-world data, rather than just performing well on the training dataset.
Let’s evaluate each option:
A) Reduce the volume of data that is used in training:
Reducing the volume of data used in training is unlikely to solve the issue and may actually make the model's performance worse. A smaller dataset might not capture the full range of variability and complexity in the data, leading to poor generalization to new, unseen data. The model could become overly simplistic or biased. Reducing the volume of data could also mean losing valuable information that would help the model to predict better in production.
B) Add hyperparameters to the model:
Adding hyperparameters could improve the model’s performance, but it doesn't directly address the issue of the model’s performance degrading in production. Hyperparameters are tuning parameters like learning rates or the number of layers in neural networks, but if the model was already trained effectively, merely adding more complexity might lead to overfitting, especially if the underlying data hasn't changed. Adjusting hyperparameters is important but would need to be paired with a more fundamental solution, like using better or more representative data.
C) Increase the volume of data that is used in training:
Increasing the volume of data used in training is a strong solution to address the issue. In production environments, data distributions may shift over time (e.g., seasonality, customer be...
Author: Liam · Last updated May 7, 2026
An ecommerce company wants to build a solution to determine customer sentiments based on written customer reviews of products.Wh...
To determine customer sentiment from written reviews, the best AWS services to meet the ecommerce company’s requirements are Amazon Comprehend and Amazon Bedrock. Here’s the reasoning for the selection:
Selected services:
1. Amazon Comprehend:
- Services and effort: Amazon Comprehend is a natural language processing (NLP) service that specializes in analyzing text for sentiment, entities, key phrases, and more. It has built-in sentiment analysis capabilities that would directly support the ecommerce company’s need to determine customer sentiment from written reviews.
- Time and cost: Comprehend is a fully managed service, meaning it reduces the time and effort required for setup and integration. It can analyze customer reviews in real-time, scaling automatically as the number of reviews increases.
- Fit: Since the task focuses specifically on determining sentiment from written text (reviews), Amazon Comprehend is the most directly applicable service. It’s cost-effective for this use case and provides the exact functionality needed.
2. Amazon Bedrock:
- Services and effort: Amazon Bedrock is a managed service that provides access to various pre-trained models, including those for NLP tasks like sentiment analysis. It allows the company to choose the most suitable foundation models for their specific use case and tailor them to their needs with minimal effort.
- Time and cost: Bedrock offers flexibility with the use of different AI models, making it highly scalable and adaptable. It provides a cost-effective way to implement advanced NLP tasks like sentiment analysis without requiring the company to build models from scratch.
Why other opt...
Author: Deepak · Last updated May 7, 2026
A company wants to use large language models (LLMs) with Amazon Bedrock to develop a chat interface for the company's product manuals. The manuals are stored as PD...
Let’s evaluate the options in terms of cost-effectiveness, effort, time, and the ability to meet the requirements of creating a chat interface for the company’s product manuals stored as PDFs:
A) Use prompt engineering to add one PDF file as context to the user prompt when the prompt is submitted to Amazon Bedrock:
This approach involves providing a single PDF file as context for the user’s query via prompt engineering. While this could work for small documents, the limitation is that only one PDF is used as context per interaction. If the user needs information from multiple product manuals, this would not scale effectively. Additionally, prompt engineering would require manual setup and customization for each document or query, which increases effort over time and might not be the most efficient solution for a large number of manuals. The cost could also increase due to frequent custom prompt generation for each request.
B) Use prompt engineering to add all the PDF files as context to the user prompt when the prompt is submitted to Amazon Bedrock:
Including all PDF files in the prompt as context would likely be inefficient and cost-prohibitive. PDFs can be large, and embedding the entirety of them in each prompt can result in high computation costs and slow performance due to the increased data size. Additionally, prompt length and input limits of the LLMs must be considered. This solution could be complex to manage as you would have to constantly reformat and adjust the prompt with all PDFs, leading to unnecessary overhead.
C) Use all the PDF documents to fine-tune a model with Amazon Bedrock. Use the fine-tuned model to ...
Author: Akash · Last updated May 7, 2026
A social media company wants to use a large language model (LLM) for content moderation. The company wants to evaluate the LLM outputs for bias and potential discrimination against specific groups or individuals.Which data...
To evaluate a large language model (LLM) for bias and potential discrimination, the company needs a data source that can provide reliable, relevant, and easily accessible content to assess the model’s outputs. Let's break down each option and its suitability based on administrative effort, cost, time, and other factors:
A) User-generated content
User-generated content is any content created by users of the platform, such as comments, posts, and messages. While this data is highly relevant to the specific content the LLM will encounter in real-world scenarios, it requires significant administrative effort to collect, clean, and label for bias evaluation. It may also involve privacy concerns and legal complexities, as personal data and sensitive content need to be handled carefully. Additionally, it can be time-consuming and costly to manually audit or assess this content for bias.
Pros: Highly relevant to the platform’s use case; reflects real-world data.
Cons: High administrative effort to clean, manage, and label data; privacy concerns and legal challenges.
B) Moderation logs
Moderation logs capture the records of content flagged or reviewed for potential policy violations, typically including user complaints, moderator decisions, and flagged content. These logs could provide insight into how the LLM might perform in content moderation tasks, and they may include instances of flagged biased or discriminatory content. However, similar to user-generated content, moderation logs can require significant effort to review for bias and may not comprehensively cover all types of discrimination or biases.
Pros: Reflects real moderation decisions and flagged content, which could help assess bias.
Cons: May not cover all instances or potential biases and may require cleaning and organizing; not as comprehensive as other options.
C) Content moderation guideli...
Author: IceDragon2023 · Last updated May 7, 2026
A company wants to use a pre-trained generative AI model to generate content for its marketing campaigns. The company needs to ensure that the generated content aligns with the company's...
The company wants to use a pre-trained generative AI model to generate content for its marketing campaigns, while ensuring that the content aligns with the company's brand voice and messaging requirements. Let's evaluate the options based on this objective.
A) Optimize the model's architecture and hyperparameters to improve the model's overall performance
Optimizing the model's architecture and hyperparameters could improve the overall performance of a generative model, but it wouldn't directly address the alignment with brand voice and messaging. Since the company is using a pre-trained model, modifying its architecture and hyperparameters would be more complex, time-consuming, and costly, and it doesn’t directly solve the problem of ensuring that the generated content matches specific brand guidelines.
- Pros:
- May improve the general performance of the model in terms of quality.
- Cons:
- Does not guarantee alignment with brand voice or messaging.
- Could require significant effort and expertise to optimize architecture and hyperparameters.
- Doesn't address the key requirement of brand alignment.
Best use case: Improving performance if the primary concern was model efficiency or output quality, but not for aligning with specific messaging needs.
B) Increase the model's complexity by adding more layers to the model's architecture
Increasing the complexity of the model (by adding more layers) may improve the model's ability to handle complex patterns, but it won't directly ensure alignment with the company's brand voice. In fact, making the model more complex could lead to longer inference times and higher resource consumption, without addressing the core need of generating content that adheres to a specific brand style.
- Pros:
- May improve the model’s capacity for more complex generation tasks.
- Cons:
- Doesn't guarantee brand alignment with the generated content.
- Increases resource requirements, which may not be necessary for the task.
- Adding complexity could lead to overfitting or irrelevant content.
Best use case: Appropriate for improving model complexity if the company had very diverse and complex content needs but not for specific brand alignment.
C) Create effective prompts that provide clear instructions and context to guide the model's gene...
Author: Olivia · Last updated May 7, 2026
A loan company is building a generative AI-based solution to offer new applicants discounts based on specific business criteria. The company wants to build and use an AI model responsibly to minimize bias that could negatively af...
In this case, the company is aiming to use an AI model responsibly to offer discounts while minimizing bias. Let's break down each option based on how well it aligns with the requirements of responsible AI usage, cost, effort, time, and impact.
A) Detect imbalances or disparities in the data
This is a crucial step to minimize bias. Imbalances or disparities in the data can lead to biased model predictions, which could unfairly affect certain groups of applicants, possibly leading to discriminatory discounting practices. Identifying and addressing these biases by analyzing the data helps ensure that the AI model provides fair outcomes for all applicants. This action is directly aligned with the company’s goal to minimize bias. The effort and time required to analyze the data for bias might be significant but are crucial for ethical model deployment.
B) Ensure that the model runs frequently
While running the model frequently may be important for timely decisions, this doesn’t directly address the issue of bias or fairness. Frequent updates or predictions don’t inherently reduce bias in the model. This option is more operational than ethical, and it doesn’t help with ensuring responsible AI usage. Hence, it’s not the best fit for this specific requirement.
C) Evaluate the model's behavior so that the company can provide transparency to stakeholders
Evaluating the model's behavior for transparency is another critical action. To ensure responsible AI use, the company must be able to explain how the model makes decisions, especially when those decisions impact customers' financial outcomes, like discounts. Transparency helps stakeholders understand how the model operates and whether there’s any potential for bias in the decision-making process. This action involves assessing and explaining the m...
Author: Elijah · Last updated May 7, 2026
A company is using an Amazon Bedrock base model to summarize documents for an internal use case. The company trained a custom model to improve the summarization quality.Which ...
Option Analysis:
A) Purchase Provisioned Throughput for the custom model:
- Reasoning: Provisioned throughput in Amazon Bedrock typically refers to setting a level of dedicated throughput for certain models to ensure performance consistency. However, simply purchasing provisioned throughput will not directly allow the company to integrate or use a custom model with Amazon Bedrock. The custom model must first be made available within Amazon Bedrock or linked appropriately, which goes beyond just throughput management.
- Use case: Provisioned throughput is relevant for ensuring that performance levels are met during high-volume use cases, but it’s not the primary step for enabling custom model use in Amazon Bedrock.
- Conclusion: Rejected.
B) Deploy the custom model in an Amazon SageMaker endpoint for real-time inference:
- Reasoning: Deploying the custom model to an Amazon SageMaker endpoint would allow real-time inference, but this would not be directly integrating it with Amazon Bedrock. Amazon Bedrock, by design, works with base models and custom models that are integrated within its ecosystem. Deploying it through SageMaker is not the typical flow for using custom models in Bedrock.
- Use case: This option is useful when the goal is to use a custom model outside the Bedrock ecosystem (direct SageMaker inference), but it’s not required for using the custom model specifically within Amazon Bedrock.
- Conclusion: Rejected.
C) Register the model with the Amazon SageMaker Model Registry:
- Reasoning: Registering the model with Amazon SageMaker’s Model Registry helps manage and track different versio...
Author: NightmareDragon2025 · Last updated May 7, 2026
A company needs to choose a model from Amazon Bedrock to use internally. The company must identify a model that generates responses in a style that the company's e...
To help the company select the right model for generating responses in a style that aligns with their employees' preferences, each option needs to be considered carefully. Let's go through the pros, cons, and potential use cases of each option:
Option A: Evaluate the models by using built-in prompt datasets
- Pros: Amazon Bedrock likely provides prompt datasets designed to evaluate the models based on general use cases. This method can save time and effort by quickly testing models against predefined scenarios.
- Cons: The built-in datasets may not perfectly reflect the company's specific internal needs or the unique style preferences of their employees. This could lead to less accurate model evaluation for the company’s exact use case.
- Use case: This option is useful for general-purpose evaluations or when the company doesn’t need highly customized responses but wants a quick comparison between available models.
Option B: Evaluate the models by using a human workforce and custom prompt datasets
- Pros: This approach offers the highest level of customization and accuracy. The company can create prompts specifically tailored to the way employees communicate, ensuring that the model generates responses that match the preferred style. In addition, a human workforce can provide qualitative feedback that is more nuanced than automated evaluation.
- Cons: This method requires a significant investment of time and effort. It could be costly depending on how large the workforce is, and might take longer to implement.
- Use case: This option is ideal when the company wants a highly personalized model response and has the time and resources to invest in fine-tuning. For example, if a company values specific tones in customer support or internal communications, this approach ensures the model is fully optimized.
Option C: Use public model leaderboards to identify the model
- Pros: Public leaderboards can give a quick overview of which m...
Author: Benjamin · Last updated May 7, 2026
A company needs to build its own large language model (LLM) based on only the company's private data. The company is concerned about the environmental effect of the training process.Whic...
To build a large language model (LLM) with a focus on minimizing environmental impact, the most appropriate choice is Amazon EC2 Trn series. Here's the reasoning:
Key factors for selecting Amazon EC2 Trn series:
- Environmental effect: The Amazon EC2 Trn series is specifically designed to optimize the energy efficiency and sustainability of machine learning workloads. These instances use specialized hardware (AWS Trainium chips) designed for training large models with a much lower carbon footprint compared to general-purpose GPUs or CPUs. They offer high performance with low power consumption, making them the most eco-friendly option for training LLMs.
- Efficiency and cost: While not as widely known as other EC2 instance types, the Trn series is highly optimized for training large models like LLMs, providing excellent performance for deep learning workloads at a lower environmental cost.
- Training LLMs: The Trn series is tailored for machine learning tasks like training large models, so it would be the ideal choice for the company's use case, ensuring not just performance but sustainability.
Why other options are rejected:
- A) Amazon EC2 C series: The C series instances are optimized for compute-heavy workloads like high-performance web servers and batch processing. While they are efficient for certain tasks, they are not designed specifically for machine lea...
Author: Kunal · Last updated May 7, 2026
A company wants to build an interactive application for children that generates new stories based on classic stories. The company wants to use Amazon Bedrock and needs to ensure that the results and topics...
The appropriate AWS service to meet the company's requirements for generating child-friendly stories with Amazon Bedrock is Guardrails for Amazon Bedrock. Here's the reasoning:
Key factors for selecting Guardrails for Amazon Bedrock:
- Services and effort: Guardrails provide a set of pre-built content filters and safety mechanisms to ensure that generated outputs are appropriate for the target audience (in this case, children). It ensures that the AI-generated stories stay within the desired boundaries (e.g., avoiding inappropriate language or themes).
- Time and cost: By using Guardrails, the company doesn't need to build complex content moderation systems themselves. Guardrails offer a scalable, automated solution, reducing both time and cost in implementing child-appropriate filters.
- Safety and appropriateness: Given the nature of the application (targeting children), it’s critical that the generated content is safe, educational, and child-friendly. Guardrails are specifically designed to help address this concern, aligning with the company's goal of generating appropriate stories.
Why other options are rejected:
- A) Amazon Rekognition: Rekognition is an image and video analysis service. While it excels at tasks like facial recognition, object detection, and text recognition in images, it is not suited for moderating or filtering text-based content. It wouldn't help in ensuring child...
Author: Suresh · Last updated May 7, 2026
A company is building an application that needs to generate synthetic data that is based on existing data.Which type of...
The company is building an application that needs to generate synthetic data based on existing data. Let's evaluate each option in the context of this requirement:
Option A: Generative adversarial network (GAN)
- Purpose: A Generative Adversarial Network (GAN) is a type of machine learning model that is specifically designed for generating synthetic data that closely resembles real data. GANs work by training two neural networks—the generator and the discriminator—in a game-like setting, where the generator creates data and the discriminator evaluates it, guiding the generator to improve.
- Reason for selection: GANs are the most suitable choice for generating synthetic data, as their core purpose is to learn the underlying distribution of the existing data and generate new, similar data. They are widely used in applications like image generation, data augmentation, and more. GANs provide the flexibility to generate high-quality synthetic data, which directly meets the company's requirement of creating data based on existing datasets.
- Effort, Time, Cost: GANs require significant training time and computational resources, but they are the most effective method for generating realistic synthetic data. The initial setup might involve substantial effort, but once trained, they can be highly efficient for generating large datasets with minimal manual intervention.
Option B: XGBoost
- Purpose: XGBoost is a popular machine learning algorithm primarily used for supervised learning tasks, such as classification and regression. It works by combining decision trees in an ensemble method to make predictions. XGBoost is not designed for generating synthetic data; instead, it excels at making predictions based on labeled datasets.
- Reason for rejection: XGBoost is no...
Author: SolarFalcon11 · Last updated May 7, 2026
A digital devices company wants to predict customer demand for memory hardware. The company does not have coding experience or knowledge of ML algorithms and needs to develop a data-driven predictive model. The company needs t...
The company needs to predict customer demand for memory hardware and doesn't have coding experience or knowledge of machine learning (ML) algorithms. This suggests that the solution should be user-friendly and allow the company to easily analyze both internal and external data without requiring coding expertise.
Let’s evaluate each option:
A) Store the data in Amazon S3. Create ML models and demand forecast predictions by using Amazon SageMaker built-in algorithms that use the data from Amazon S3.
- Rejection Reasoning: While storing the data in Amazon S3 and using SageMaker built-in algorithms is a valid approach, it requires a certain level of expertise in machine learning and coding to prepare the data, choose the right algorithms, and interpret the results. Since the company does not have coding experience, this option may not be the best fit for them as it would likely require more technical effort.
B) Import the data into Amazon SageMaker Data Wrangler. Create ML models and demand forecast predictions by using SageMaker built-in algorithms.
- Rejection Reasoning: Amazon SageMaker Data Wrangler is a powerful tool for data preparation, and it simplifies the process of importing and transforming data. However, building models and making predictions with SageMaker built-in algorithms still involves some complexity in terms of understanding the algorithms and model deployment. Although Data Wrangler helps with data preparation, the overall solution may still require more technical know-how than the company possesses.
C) Import the data into Amazon SageMaker Data Wrangler. Build ML models and demand forecast predictions by using an Amazon Pers...
Author: Leo · Last updated May 7, 2026
A company has installed a security camera. The company uses an ML model to evaluate the security camera footage for potential thefts. The company has discovered that the model disproportionately flags people wh...
To address the type of bias affecting the model’s output in this scenario, let's analyze the options based on the description that the model disproportionately flags people from a specific ethnic group. The goal is to identify the type of bias that directly leads to this situation. Each option will be examined in the context of the problem.
Option A: Measurement Bias
- Explanation: Measurement bias occurs when there is an error in how the data is measured or collected, leading to incorrect or skewed results. This could happen if, for example, the camera's resolution, angle, or environmental conditions distort the features of individuals in a way that disproportionately affects certain groups.
- Relevance: While measurement bias can cause issues in how data is captured, it doesn't seem to directly address the disproportionate flagging of one specific ethnic group. The problem described seems more related to the model's behavior rather than the raw data quality or capture process itself.
- Use case: Measurement bias would apply if the camera setup itself is flawed, leading to errors in the raw footage, but it's less relevant in this case where the issue seems to be with the model’s interpretation of data.
Option B: Sampling Bias
- Explanation: Sampling bias occurs when the data used to train the model is not representative of the real-world population, meaning that certain groups are either overrepresented or underrepresented in the training data. In this case, if the model was trained on data that disproportionately represented one ethnic group (or did not include enough diversity), the model may learn patterns that result in biased flagging of individuals from that group.
- Relevance: This option is highly relevant because it directly explains the issue. If the model was trained on a dataset that is not representative of the real-world distribution of ethnicities or over-represented certain groups, the model would develop biased behavior, flagging certain individuals more frequently based on ethnicity.
- Use case: Sampling bias is a strong contender when the model’s training data is skew...
Author: Emma · Last updated May 7, 2026
A company is building a customer service chatbot. The company wants the chatbot to improve its responses by learning from past interactions and online resources....
To build a customer service chatbot that can improve its responses over time by learning from past interactions and online resources, the company needs a learning strategy that allows for continuous adaptation and refinement. Let’s evaluate the options in detail:
A) Supervised learning with a manually curated dataset of good responses and bad responses
In supervised learning, the model learns from a labeled dataset, which includes pairs of inputs (e.g., customer inquiries) and corresponding outputs (e.g., good and bad responses). This approach can work well for training the chatbot initially with examples of both effective and ineffective responses, but it doesn’t inherently support the idea of continuous self-improvement. The model can only improve based on the fixed dataset it was trained on, so without constant manual updates and curation of new good and bad responses, this approach would not support dynamic learning from ongoing interactions.
Pros: Good for initial training; can improve with high-quality curated datasets.
Cons: Lacks continuous self-improvement and adaptation; requires manual effort to update the dataset.
B) Reinforcement learning with rewards for positive customer feedback
Reinforcement learning (RL) enables the chatbot to learn by interacting with its environment (e.g., customers) and receiving feedback. The model can adjust its behavior based on rewards (positive feedback) or penalties (negative feedback) for specific actions. This strategy is highly suited for a chatbot that wants to learn from its past interactions and improve over time. The feedback mechanism allows for dynamic learning, and the model can continuously adapt its responses as it receives more interactions. This is a true self-improvement capability, as the chatbot learns from both successes and mistakes.
Pros: Enables continuous, dynamic learning based on real interactions; can self-improve over time.
Cons: Requires significant setup and management to ensure that the feedback system is properly implemented and that t...
Author: Amira · Last updated May 7, 2026
An AI practitioner has built a deep learning model to classify the types of materials in images. The AI practitioner now wants to measure the model performance.Which me...
To evaluate the performance of a deep learning model that classifies the types of materials in images, we need to choose an appropriate metric based on the task (classification) and the evaluation goals.
A) Confusion matrix
A confusion matrix is a powerful tool for evaluating classification models, especially when there are multiple classes. It provides insight into how well the model distinguishes between different types of materials, by showing the counts of true positive, true negative, false positive, and false negative predictions for each class. This can help identify which material types the model is confusing with others and assess the overall performance in terms of precision, recall, and F1-score.
- Pros:
- Perfect for classification tasks, particularly multi-class classification.
- Helps in understanding where the model is making errors and which classes it struggles with.
- Provides detailed metrics like precision, recall, and F1-score, which are very informative in classification.
- Cons:
- Does not give a single numeric measure of overall performance by itself (but is used as a basis for other metrics).
Best use case: Evaluating classification tasks, especially when the goal is to analyze the specific performance of the model for each class.
B) Correlation matrix
A correlation matrix is used to measure the correlation between different variables. It is typically used for regression tasks or when analyzing relationships between continuous variables. In a classification scenario, this wouldn't be as useful because the target variable (material types) is categorical, not continuous.
- Pros:
- Useful for identifying correlations between continuous variables.
- Cons:
- Not suitable for classification tasks, especially multi-class classification.
- Doesn’t directly evaluate model performance.
Best use case: Understanding the relationships ...
Author: Noah · Last updated May 7, 2026
A company has built a chatbot that can respond to natural language questions with images. The company wants to ensure that the chatbot does not return inapprop...
Let's analyze each option based on the requirement: ensuring that the chatbot does not return inappropriate or unwanted images. We'll consider the effectiveness, effort, time, cost, and suitability of each solution.
Option A: Implement moderation APIs
Moderation APIs are designed to automatically filter and flag inappropriate content, including images, based on predefined criteria such as violence, nudity, or hate speech. These APIs can be integrated into the chatbot system, ensuring that any images returned by the chatbot are checked before being delivered to the user. This is a straightforward solution that reduces the effort, time, and cost associated with manually curating content. It also scales easily with minimal changes required in the chatbot's workflow.
Selected reason: Moderation APIs provide an effective and automated way to filter out inappropriate or unwanted images. They are specifically designed for this kind of content moderation task and are easy to implement, making them the ideal solution.
Option B: Retrain the model with a general public dataset
Retraining the model with a more diverse dataset may help to some extent in improving the quality of the model's outputs. However, this would not directly address the issue of ensuring that the model returns appropriate or unwanted images. Retraining can be resource-intensive, time-consuming, and expensive, and may not guarantee that the model will completely avoid generating inappropriate content, as even with a well-balanced dataset, some risks might still remain.
Rejection reason: While retraining could potentially reduce some inappropriate outputs, it doesn't specifically address the need for content moderation and may not be as reliable or efficient as using dedicated moderation APIs. It's a more complex and expensive approach.
...
Author: Amira · Last updated May 7, 2026
An AI practitioner is using an Amazon Bedrock base model to summarize session chats from the customer service department. The AI practitioner wants to store invocation logs to moni...
To monitor model input and output data effectively when using Amazon Bedrock for summarizing session chats from the customer service department, the AI practitioner needs a strategy to store invocation logs and track both model requests and responses. Let's analyze each of the options provided:
A) Configure AWS CloudTrail as the logs destination for the model.
- Rejection Reasoning: AWS CloudTrail is primarily used for logging API activity across AWS services, which helps in auditing and tracking who did what and when. While it’s useful for monitoring AWS API calls and security-related information, it is not designed specifically for logging model inputs and outputs. CloudTrail won’t capture detailed information like the content of the model invocations or responses, making it unsuitable for tracking the specific data of model interactions.
B) Enable invocation logging in Amazon Bedrock.
- Selection Reasoning: Amazon Bedrock offers invocation logging as part of its functionality, which is specifically designed to track model inputs and outputs. By enabling invocation logging directly within Amazon Bedrock, the AI practitioner can ensure that the inputs and outputs of each model invocation are stored for monitoring, analysis, and auditing purposes. This option is optimized for the use case of tracking model interactions in real-time, and it will provide the necessary logs to assess model performance or review output from cust...
Author: Liam123 · Last updated May 7, 2026
A company is building an ML model to analyze archived data. The company must perform inference on large datasets that are multiple GBs in size. The company does not need to access the model predicti...
To determine the best solution for performing inference on large datasets that are multiple GBs in size, where immediate access to the model predictions is not required, we need to evaluate the available Amazon SageMaker inference options in the context of the requirements.
Option A: Batch transform
- Description: Batch Transform is an Amazon SageMaker feature that allows you to process large datasets in bulk. It is ideal for inference on large datasets where the model predictions can be processed asynchronously.
- Key factors:
- Large datasets: Batch Transform is specifically designed to handle large datasets efficiently, making it suitable for multi-GB datasets.
- Inference timing: Since the company does not require immediate access to the predictions, Batch Transform fits well because it processes the data in batches and stores the results in an S3 bucket, allowing you to access them when ready.
- Effort & time: The process is handled in a batch manner, and although it may take time depending on the dataset size, it's efficient for large-scale inference.
- Cost: You only pay for the compute resources used during the inference job, and the costs are relatively predictable for batch processing.
- Why it’s suitable: Batch Transform is designed for situations where you need to process large volumes of data asynchronously and do not need immediate access to predictions.
Option B: Real-time inference
- Description: Real-time inference in Amazon SageMaker is used for low-latency predictions, where the model needs to provide results immediately after receiving input.
- Key factors:
- Large datasets: Real-time inference is typically not ideal for multi-GB datasets, as it is designed for smaller, quicker requests. Handling large datasets in real-time would require splitting the data into smaller parts, which adds complexity.
- Inference timing: Since real-time inference requires the model to respond instantly, it’s not suitable for scenarios where the predictions don't need to be accessed immediately.
- Cost: Real-time inference incurs charges based on the number of requests and the compute resources used. If the dataset is large and inference isn't required immediately, this could lead to inefficient usage and higher costs.
- Why it’s not suitable: This is not a good fit for large datasets that don't need to be processed in real-time.
Option C: Serverless inference
- Description: Serverless inference allows you to run models ...
Author: Akash · Last updated May 7, 2026
Which term describes the numerical representations of real-world objects and concepts that AI and natural language processing (NLP) m...
Option Analysis:
A) Embeddings:
- Reasoning: Embeddings are numerical representations of real-world objects, words, or concepts in a lower-dimensional space that AI and NLP models use to improve understanding. They capture semantic relationships between words and can represent complex concepts by encoding the meaning and context of terms. Embeddings are essential in AI and NLP because they help models interpret and process textual data more effectively.
- Use case: Embeddings are specifically used to enhance understanding in tasks like text classification, sentiment analysis, machine translation, and more. They are a key part of most modern NLP models like BERT and GPT.
- Conclusion: Selected.
B) Tokens:
- Reasoning: Tokens refer to the individual pieces of text that are processed by an NLP model (e.g., words, characters, or subwords). While tokens are essential for NLP tasks, they represent the raw input rather than the numerical representations that improve understanding. Tokens themselves are not used to encode semantic meaning, which is the core purpose of embeddings.
- Use case: Tokens are used as part of preprocessing in NLP, but they are not the direct answer to the question of numerical representations used for improving understanding.
- Conclusion: Rejected.
C) Models:
- Reasoning: Models are the algorithms or systems used to process and analyze data. While models use embeddings as part of their architecture to improve understanding, models...
Author: Ishaan · Last updated May 7, 2026
A research company implemented a chatbot by using a foundation model (FM) from Amazon Bedrock. The chatbot searches for answers to questions from a large database of research papers.After multiple prompt engineering attempts, the company notices that the FM is performing poorly b...
In this scenario, the research company is using a foundation model (FM) from Amazon Bedrock to implement a chatbot, but the FM is struggling with complex scientific terms found in the research papers. The company wants to improve the chatbot's performance to better handle these terms. Let’s evaluate each of the options:
Option A: Use few-shot prompting to define how the FM can answer the questions
- Explanation: Few-shot prompting involves providing a small number of examples within the prompt to guide the model on how to answer specific questions. This method can be effective when you need the model to understand how to respond to specific types of questions or tasks, but it may not directly solve the issue of understanding complex scientific terms.
- Relevance: While few-shot prompting could help guide the model’s behavior, it doesn't directly address the core issue, which is the model’s difficulty with complex scientific terminology. Few-shot prompting would help the model structure its responses more effectively but won't necessarily improve its understanding of specialized language.
- Use case: This option is useful if the problem is related to response formatting or specific types of questions, but it won't help with the model's understanding of complex scientific terms in the research papers.
Option B: Use domain adaptation fine-tuning to adapt the FM to complex scientific terms
- Explanation: Domain adaptation fine-tuning involves training the foundation model on a domain-specific dataset to help the model better understand the language and terminology specific to that domain. In this case, fine-tuning the FM on research papers containing complex scientific terms will help the model adapt to those terms and improve its understanding and performance.
- Relevance: This option is highly relevant because it directly addresses the issue of the model’s poor performance with scientific terms. By fine-tuning the FM on research papers that use the terminology the model is struggling with, the model will learn how to better handle such terms and improve its performance.
- Use case: Domain adaptation fine-tuning is ideal in this scenario because it allows the company to adapt the model to the specific language of their research papers, improving the chat...
Author: Isabella1 · Last updated May 7, 2026
A company wants to use a large language model (LLM) on Amazon Bedrock for sentiment analysis. The company needs the LLM to produce more consistent responses to the same input prompt.Which adjus...
To make the responses from the large language model (LLM) on Amazon Bedrock more consistent for the same input prompt, the company should decrease the temperature value. Here's the reasoning:
Key factors for selecting Decrease the temperature value:
- Temperature parameter: In LLMs, the temperature setting controls the randomness of the model's output. A higher temperature leads to more diverse, creative, and potentially unpredictable responses, while a lower temperature generates more deterministic and consistent responses.
- Consistency: By decreasing the temperature value, the model becomes more likely to produce the same output for identical input prompts. This is exactly what the company needs for ensuring consistent sentiment analysis results.
- Fit: In sentiment analysis, consistency is important as the company wants reliable, repeatable results. Lowering the temperature ensures that the model doesn't introduce unnecessary randomness, which is critical in scenarios where accuracy and reliability are key.
Why other options are rejected:
- B) Increase the temperature value: Increasing the temperature value makes the model’s responses more varied and creative, which can lead to inconsistent answers for the same input. This would not help with achieving consistency in sentiment analysis, where reliable output i...
Author: Aarav · Last updated May 7, 2026
A company wants to develop a large language model (LLM) application by using Amazon Bedrock and customer data that is uploaded to Amazon S3. The company's security policy states that each team can acc...
To meet the company's requirements of using Amazon Bedrock to develop a large language model (LLM) application while adhering to the security policy that restricts teams to accessing only their own customer data in Amazon S3, let's evaluate each option based on factors like security, ease of implementation, scalability, and adherence to the policy.
A) Create an Amazon Bedrock custom service role for each team that has access to only the team's customer data
- Benefits: This option involves creating a dedicated service role for each team, restricting their access to only the customer data belonging to their team. It directly aligns with the security policy, ensuring that teams only access the data they are permitted to. This solution provides granular access control, which is ideal for compliance.
- Drawback: This approach may increase the administrative overhead, as creating and managing multiple service roles could become complex as the number of teams grows. However, this is outweighed by the strong security benefits.
- Use Case: This is the most direct solution to meet the policy requirements, especially when fine-grained access control is crucial and the number of teams is manageable.
B) Create a custom service role that has Amazon S3 access. Ask teams to specify the customer name on each Amazon Bedrock request
- Benefits: This option simplifies access control by using a single service role for Amazon S3 access, with teams specifying the customer name during each request.
- Drawback: The security of this solution is weak because it relies on the teams to correctly specify the customer name. This could lead to potential mistakes or misconfigurations, which is a risk for data privacy and compliance. It also doesn't provide automatic enforcement of the policy at the service level, and it adds manual responsibility for ensuring correct data access.
- Use Case: This option could be used in low-risk scenarios where teams are trusted to manage data access correctly, but it does not offer strong security enforcement, making it less ideal for this case.
C) Redact personal data in Amazon S3. Update the S3 bucket policy to allow team access to customer data
- Ben...
Author: Sofia · Last updated May 7, 2026
A medical company deployed a disease detection model on Amazon Bedrock. To comply with privacy policies, the company wants to prevent the model from including personal patient information in its responses. The company also wa...
Let’s analyze the options in relation to the requirements of ensuring privacy and compliance for the disease detection model:
A) Use Amazon Macie to scan the model's output for sensitive data and set up alerts for potential violations:
Amazon Macie is a service that helps identify and protect sensitive data, such as personally identifiable information (PII), in Amazon S3 buckets. While Macie can be useful for detecting sensitive data, it is not directly integrated with Amazon Bedrock, and the task here is related to scanning the model’s outputs, not data stored in S3. Using Macie would require additional infrastructure and may not be as seamless for this use case, especially for real-time detection of model output. This solution might be too indirect and require additional effort in integration.
B) Configure AWS CloudTrail to monitor the model's responses and create alerts for any detected personal information:
AWS CloudTrail logs API calls and provides visibility into user activity, but it doesn’t provide a direct mechanism to scan or analyze model outputs for sensitive information. It is great for tracking activity, but it doesn't help in scanning or filtering content like PII. For this use case, CloudTrail would not be effective because it doesn’t actively scan the data produced by the model in real time. It’s more about audit logs and tracking access, not about ensuring compliance with privacy policies by analyzing model responses.
C) Use Guardrails for Amazon Bedrock to filter content. Set up Amazon CloudW...
Author: Zara · Last updated May 7, 2026
An education provider is building a question and answer application that uses a generative AI model to explain complex concepts. The education provider wants to automatically change the style of the model response depending on who is asking the question. The education provider will give the model t...
The education provider wants to dynamically adjust the style of the model's response based on the age range of the user asking the question. Let's evaluate each option with respect to the least implementation effort, considering factors like ease of setup, flexibility, and maintenance:
A) Fine-tune the model by using additional training data that is representative of the various age ranges that the application will support:
- Effort: Fine-tuning involves significant effort in collecting age-specific data, preparing the training pipeline, retraining the model, and validating the results. This option requires substantial investment in terms of time and resources to create datasets, retrain the model, and manage the ongoing maintenance of these different versions.
- Why rejected: Although it can work, it demands considerable ongoing effort and complexity. The cost and effort associated with this approach are higher than the other solutions.
B) Add a role description to the prompt context that instructs the model of the age range that the response should target:
- Effort: This is a relatively low-effort solution. By simply modifying the prompt to include an instruction about the age range, the model can adjust its response style accordingly (e.g., "explain this to a 10-year-old" or "explain this to a college student"). It leverages the flexibility of prompt engineering to steer the model's responses without needing to retrain or fine-tune.
- Why selected: This approach directly addresses the requirement of modifying the response style with minimal effort and complexity. It is adaptable and doesn’t require heavy resources or significant reworking of the underlying model.
C) Use chain-of-thought reasoning to deduce the correct style and complex...
Author: James · Last updated May 7, 2026
Which strategy evaluates the accuracy of a foundation model (FM) that is used in image classificatio...
When evaluating the accuracy of a foundation model (FM) used in image classification tasks, it is essential to focus on how well the model performs at classifying images. Here's a breakdown of each option:
A) Calculate the total cost of resources used by the model:
- Effort & Relevance: This option assesses the resource consumption (e.g., computational cost, memory usage, etc.), but it doesn't measure how accurately the model performs in its image classification task. Cost-related factors may be important for deployment or scalability but are not a metric for model accuracy.
- Why rejected: While useful for operational insights, this does not provide any insight into the model's performance or accuracy in classifying images.
B) Measure the model's accuracy against a predefined benchmark dataset:
- Effort & Relevance: This strategy directly measures the model's performance by comparing its predictions against a known dataset where the correct answers (labels) are predefined. This is a standard and effective method for evaluating accuracy in classification tasks, including image classification. By calculating metrics like accuracy, precision, recall, and F1 score, the model's performance is clearly assessed.
- Why selected: This is the most common and accurate approach to evaluate the performance of a model in image classification tasks. Benchmark datasets (e.g., ImageNet, CIFAR-10) are widely used for this purpose.
C) Count the number of layers in the neural network:
- Effort & Relevance: Whil...
Author: Jack · Last updated May 7, 2026
An accounting firm wants to implement a large language model (LLM) to automate document processing. The firm must proceed responsibly to avoid potential harms.What ...
In this scenario, the accounting firm needs to implement a large language model (LLM) for document processing while ensuring the model behaves responsibly and avoids potential harms. The firm must consider fairness, bias mitigation, model robustness, and ethical deployment.
Option Analysis:
A) Include fairness metrics for model evaluation.
- Pros: Evaluating the model with fairness metrics ensures that it performs equitably across various demographic groups or use cases. This is crucial in document processing, where biased outputs could have legal or financial consequences. Fairness metrics help identify and mitigate potential harms like discrimination or unequal treatment based on sensitive factors.
- Cons: Implementing fairness metrics can add some complexity to the evaluation process and may require additional tools or frameworks to measure fairness. This may increase effort and time, but the benefits in reducing harm are worth it.
- When to use: This should be used as a part of responsible AI practices in any application, particularly when dealing with legal, financial, or sensitive data, as in the case of the accounting firm.
B) Adjust the temperature parameter of the model.
- Pros: Adjusting the temperature in a model controls the randomness of the model’s responses. A lower temperature makes the model’s output more deterministic, while a higher temperature makes it more creative or varied.
- Cons: While adjusting the temperature can affect the model's creativity, it does not directly address responsible deployment, fairness, or ethical concerns. It is more about controlling the output style, not mitigating potential harms such as bias or unethical decision-making in document processing.
- When to use: Adjusting the temperature could be useful for fine-tuning the model's output for specific tasks (e.g., balancing creativity and precision), but it is not central to ensuring the model is deployed responsibly.
C) Modify the training data to mitigate bias.
- Pros: Bias in the training data can lead to biased outputs, which is a significant concern in many applications, especially in finance and legal documents. Modifying the training data to reduce bias ensures that the model outputs are more fair and representative, reducing potential harm in decision-making.
- Cons: This step may involve significant effort, as it requires carefully curating the data to avoid introducing biases or un...
Author: Noah Williams · Last updated May 7, 2026
A company is building an ML model. The company collected new data and analyzed the data by creating a correlation matrix, calculating statistics, and visualizing...
The company is currently in the Exploratory Data Analysis (EDA) stage. Here's the reasoning:
Key factors for selecting EDA:
- Services and effort: EDA involves analyzing and understanding the data before any model-building begins. This can include visualizations, calculating basic statistics, and examining relationships between variables (like using a correlation matrix). It's more about gaining insights into the data's structure.
- Time and cost: This phase typically requires significant time for investigation and understanding, but it is a crucial part of the model pipeline as it guides subsequent decisions.
- Focus on Data: The company is focusing on data exploration and understanding the relationships within the data (as indicated by their use of a correlation matrix and visualizations).
Why other options are rejected:
- A) Data pre-processing: Data pre-processing usually comes before EDA, where tasks like cleaning data (removing missing values, handling outliers), normalization, or transformation take place. EDA is more about exploration and gaining insights, while pre-processing focuses on preparing the da...
Author: Sofia · Last updated May 7, 2026
A company has documents that are missing some words because of a database error. The company wants to build an ML model that can suggest potential words to ...
In this scenario, the company needs to suggest potential words to fill in missing text in documents. This suggests the need for a model that can understand the context of the surrounding text and generate or predict missing words accordingly.
Option Analysis:
A) Topic Modeling:
- Reasoning: Topic modeling techniques (such as Latent Dirichlet Allocation - LDA) are used to identify the underlying themes or topics within a collection of documents. While it can give a high-level understanding of document themes, it does not focus on predicting or filling in missing words in the context of the text. This is not ideal for this specific task.
- Use case: Topic modeling would be useful for categorizing documents based on themes, but not for filling in missing text.
- Conclusion: Rejected.
B) Clustering Models:
- Reasoning: Clustering models group similar documents or text together based on features. However, these models do not directly address the problem of predicting missing words. They are used for grouping data, but not for text completion.
- Use case: Clustering models might be helpful for organizing documents or detecting similar content, but they would not be able to suggest words for missing text.
- Conclusion: Rejected.
C) Prescriptive ML Models:
- Reasoning: Prescriptive ML models provide recommendations for actions based on data and insights, but they are often more concerned with making decisions or suggest...
Author: Liam123 · Last updated May 7, 2026
A company wants to display the total sales for its top-selling products across various retail locations in the past 12 months.Which AWS solutio...
The company wants to automate the generation of graphs showing the total sales for its top-selling products across various retail locations over the past 12 months. The goal here is to visualize this sales data efficiently with minimal manual effort.
Let's evaluate each option:
Option A: Amazon Q in Amazon EC2
- Purpose: Amazon EC2 is an IaaS (Infrastructure as a Service) that provides virtual machines (instances) for running applications. While EC2 can be used to run custom applications (including data analytics tools or custom dashboards), it requires manual setup, management, and scaling of infrastructure. Using EC2 would not be an optimal solution for automating graph generation since the company would need to configure and manage the application stack.
- Reason for rejection: EC2 provides infrastructure but lacks automation and is not tailored for the direct purpose of automating the generation of visual data insights such as graphs. It would require more effort and resources compared to other options.
Option B: Amazon Q Developer
- Purpose: Amazon Q Developer is a tool designed to help developers build generative AI applications. While it can help create applications and automate some tasks, it isn't specifically focused on data visualization or graph generation for sales data.
- Reason for rejection: While generative AI can be useful for some automated tasks, this option is not aligned with the specific need for visualizing sales data in the form of graphs. It would be an over-engineered solution for this use case, where a straightforward data visualization tool is better suited.
Option C: Amazon Q in Amazon QuickSight
- Purpose: Amazon QuickSight is a scalable business intelligence (BI) service that allows users to create visualizations, reports, and dashboards from their data. Amazon Q is a feature within QuickSight that allows users to query data using natural language (NLQ, or natural language queries) an...
