Question 1

A financial services enterprise is standardizing on Amazon SageMaker AI for feature engineering and model lifecycle, while Amazon Fraud Detector scores real-time card transactions using model endpoints hosted on SageMaker AI. Data scientists routinely train and tune machine learning (ML) models on highly confidential payment data inside a dedicated VPC.  The security team must prevent data exfiltration paths from notebooks, training jobs, and hosted endpoints, especially through public internet routes or unauthorized client locations, without breaking internal access for analysts on the corporate network.  Which solution will meet this requirement while minimizing data egress from SageMaker AI? (Select THREE.)

Accepted Answer

Run training jobs and models with network isolation enabled (EnableNetworkIsolation=true)., Enforce IAM policies that restrict notebook presigned URLs to approved corporate IP ranges using aws:SourceIp and related conditions., Route all SageMaker AI API/Runtime calls through VPC interface endpoints (AWS PrivateLink).

Question 2

A publishing company is developing an internal generative AI platform to automate the creation of research article drafts and editorial summaries. The company uses Amazon Comprehend for text preprocessing and data cleaning, and Amazon Bedrock Agents to automate content retrieval, citation validation, and orchestration of large language model (LLM) tasks across the editorial workflow. An AI developer has recently built and fine-tuned a large language model (LLM) outside of Amazon SageMaker AI and stored the model artifacts in an Amazon S3 bucket for internal use.  The AI developer wants to make the model available to the data specialist team through SageMaker Canvas, allowing the data specialists to experiment with the model’s text-generation capabilities through a no-code interface. The AI developer and the data specialist team belong to the same SageMaker AI domain, and the AI developer must ensure the model is properly registered and accessible within SageMaker AI.  Which combination of steps must be taken for the AI developer to enable SageMaker Canvas access to the model? (Select TWO.)

Accepted Answer

The data specialist team must be granted the necessary permissions to access the S3 bucket where the model artifacts are stored., The AI developer must register the model in the SageMaker Model Registry to enable the data specialist team's access via SageMaker Canvas.

Question 3

A digital health startup is developing a compliance pipeline to process patient documents, including scanned intake forms, handwritten prescriptions, and clinical notes. The organization uses Amazon Textract to extract text and structured data from the scanned documents and Amazon Rekognition to detect and blur faces appearing in uploaded medical images.  After text extraction, the company must identify and remove personally identifiable information (PII) and protected health information (PHI) before storing the content in Amazon S3 for downstream analytics and large language model training. The compliance team must ensure that the process meets HIPAA and data privacy standards while preserving relevant medical details such as medications, diagnoses, and treatment plans for future analysis.  Which services should be used to automatically detect and redact both PII and PHI entities from the extracted clinical text before it is stored or processed further? (Select TWO.)

Accepted Answer

Use Amazon Comprehend to detect and redact personally identifiable information (PII) from the text extracted.

Question 4

A multinational publishing company is developing an internal AI-powered assistant to help employees retrieve, summarize, and analyze research documents. The company’s data-science team already relies on Amazon SageMaker Ground Truth to label training data and improve metadata accuracy.  The next development phase involves integrating AI agents in Amazon Bedrock that can autonomously answer questions, summarize lengthy documents, and recommend related content. To achieve this, the team requires a mechanism that can represent both user queries and document content as semantic embeddings that capture contextual relationships between concepts and terms.  The solution must be fully managed, scalable, and able to integrate easily with existing AWS services to support downstream retrieval-augmented generation (RAG) workflows.  Which approach best satisfies these requirements?

Accepted Answer

Leverage Amazon Titan Text Embeddings in Bedrock to convert text into semantic vectors and store them in Amazon OpenSearch Service for context-aware retrieval and reasoning by AI agents.

Question 5

A global e-commerce company uses Amazon Bedrock with the Amazon Titan foundation model (FM) to power a multilingual customer support chatbot. In addition to Titan FM, the company also uses Amazon Comprehend to perform entity recognition and sentiment analysis on incoming customer messages. However, the support chatbot often provides generic answers because it lacks access to the company’s proprietary order management and product documentation data stored in Amazon S3 and an internal database.  The AI developer wants to enhance the chatbot’s responses with retrieval-augmented generation (RAG) so that Titan FM can generate more accurate and contextually relevant replies based on real-time, private company data, without retraining the model.  Which of the following options will satisfy this requirement?

Accepted Answer

Set up a Bedrock knowledge base and integrate it with the company’s private data sources.

Question 6

A travel company is developing a virtual assistant using Amazon Lex to help customers find vacation packages based on themes such as “relaxation,” “adventure,” and “culture.” The chatbot uses an AWS Lambda function to query an Amazon DynamoDB table that stores package details by category.  During testing, a Generative AI Developer observes that the chatbot sometimes fails to recognize user inputs such as “thrill-seeking,” “sightseeing,” or “chill,” even though these terms correspond to existing categories. The company is exploring Amazon Titan models to enhance future natural language understanding and embedding capabilities but requires an immediate solution that does not involve modifying the Lambda function or database.  Which action should the Generative AI Developer take to improve the chatbot’s ability to recognize these user inputs?

Accepted Answer

Define the unrecognized words as synonyms linked to current enumeration values in the custom slot type.

Question 7

A financial services organization is developing a document classification model to detect fraudulent claims from scanned forms. The team uses Amazon Textract to extract text and structured data from thousands of claim documents, then builds a training dataset in Amazon SageMaker AI. During model evaluation, the data scientists notice that the model performs very well on “legitimate claim” documents but frequently misclassifies “fraudulent claim” samples.  To understand the cause, the team uses SageMaker Clarify and observes that the pretraining bias analysis reveals a significant skew in the dataset. The team realizes this uneven class distribution leads to biased predictions and poor generalization to minority classes.  What issue is most likely causing the model’s poor performance on fraudulent claim detection?

Accepted Answer

The issue is due to class imbalance (CI) in the training dataset, where the minority class has too few samples, causing the model to learn biased decision boundaries.

Question 8

A cloud development team is using Amazon Q Developer, a generative AI-powered conversational assistant, to accelerate work on AWS applications. The team wants Q Developer to deliver more dynamic and contextually relevant responses by accessing live data from databases and external APIs that provide up-to-date information.  To ensure that Q Developer receives real-time and dynamic contextual data during conversations, the team has decided to use the Model Context Protocol (MCP). A solution is needed that can easily connect to these data sources and enable low-latency access, providing the most relevant information to the assistant during interactions.  Which solution will meet this requirement?

Accepted Answer

Utilize Amazon Q Developer CLI with MCP to connect to the databases and the external APIs.

Question 9

A national utility company manages hourly power consumption data for 150 types of smart meters deployed across multiple regions. Over the past 30 years, the company has collected a large volume of historical usage data enriched with weather metrics, regional energy prices, and maintenance logs.  The data science team currently uses Amazon SageMaker AI Data Wrangler to perform feature engineering and data preprocessing, and SageMaker AI Canvas (a no-code ML service) to generate baseline forecasts for monthly energy demand. However, the team now plans to build a custom machine learning (ML) model that can forecast future meter-usage patterns across all meter types.  Which option satisfies the requirement with the LEAST operational overhead?

Accepted Answer

Build a single global forecasting model using the SageMaker AI DeepAR forecasting algorithm trained on all meter types to capture shared consumption patterns.

Question 10

An enterprise is currently leveraging Amazon SageMaker Feature Store for cross-model reusable feature lineage and SageMaker Clarify for model bias monitoring pre-deployment. A new model is now planned to be deployed using a SageMaker AI endpoint. The enterprise security team mandates that all inference traffic must remain inside private subnets, so the SageMaker AI endpoint must use a VPC configuration with no public internet route. Expected request payload sizes are consistently 6 MB to 11 MB, and model inference execution time during operational peak can take 18–22 minutes per request. The finance engineering team also requires that costs must remain minimal while still supporting request/response style inference.  Which solution is the most suitable approach to satisfy this requirement?

Accepted Answer

Deploy a SageMaker asynchronous endpoint inside private subnets and include VPC configuration parameters during endpoint creation.

Question 11

An organization is developing a machine learning-based fraud detection system to process real-time transactions. The system leverages multiple APIs for generating embeddings from transaction descriptions, which are then used to assess the likelihood of fraud. These APIs are integrated with Amazon Comprehend for natural language processing tasks and Amazon SageMaker AI to handle advanced machine learning tasks for analyzing transaction patterns and building custom fraud detection models.  The organization’s security policy requires rotating API tokens every 3 months to reduce exposure risks. The solution must automate token rotation, maintain strong security, and integrate smoothly with existing AWS services. It must also ensure secure token storage, automatic updates within applications, and continuous operation without downtime.  Which solution will best address these requirements?

Accepted Answer

Use AWS Secrets Manager to store the tokens and rely on an AWS Lambda function to perform the rotation process.

Question 12

A social media platform is developing an ML model to moderate user-generated content. The platform stores large volumes of textual and document data in Amazon S3. It uses Amazon Textract to extract text from uploaded PDFs or images, and Amazon SageMaker AI to train custom language models that classify comments based on sentiment, context, and topic relevance.  The moderation team aims to enhance safety by detecting toxic or harmful language in real-time, including hate speech, harassment, and explicit threats. The solution must integrate seamlessly with the existing SageMaker AI inference pipeline, handle high throughput, and provide confidence scores for each classification, allowing content flagged as toxic to be reviewed or blocked automatically.  Which of the solutions provides a managed solution for detecting toxicity in text to support this ML pipeline?

Accepted Answer

Utilize Amazon Comprehend toxicity detection to identify abusive or harmful language in text.

Question 13

A global publishing company manages a large repository of multimedia content, including PDFs, images, audio files, and video recordings from news articles, interviews, and webinars. The company seeks to automate the extraction of insights such as key topics and relevant entities to help journalists and editors quickly access information. The company aims to build a generative AI-powered research assistant capable of answering natural language queries like, “Which article mentioned climate adaptation in Southeast Asia and included aerial images?”  The company plans to use Amazon Bedrock Data Automation (BDA) for processing unstructured media and extracting structured insights, and Amazon SageMaker AI to host the generative AI model for context‑aware response generation.  Which approach will best meet the requirements?

Accepted Answer

Leverage Bedrock Data Automation (BDA) to process documents, images, audio, and video, index the results in Bedrock Knowledge Bases for semantic search, and then input the retrieved context into a foundation model via SageMaker AI for response generation.

Question 14

A global online publishing company is developing a generative AI-based content recommendation system to suggest relevant tags for articles. The raw data for the articles, stored in JSON format, is housed in an Amazon S3 bucket. The data scientist is using the Amazon SageMaker Neural Topic Model (NTM) algorithm to build a model that can generate the most appropriate tags for each article. To enhance the model’s performance, Amazon Titan is being utilized to improve natural language understanding and generate contextual insights from the article content, leveraging Titan’s powerful pre-trained models for tag suggestions.  During the evaluation phase, the model’s recommendations include common stopwords such as “and,” “of,” and “in,” along with some rare words that appear only in a few specific articles. After working with the content team to review the model’s output, the data scientist determines that while the rare words are unusual, the words could still be valuable for generating more relevant tags. The challenge is to find a way to exclude the stopwords from the model’s tag predictions while still retaining the rare but meaningful words.  How can the model be refined to exclude stopwords and retain rare words in tags?

Accepted Answer

Employ the CountVectorizer function from scikit-learn to preprocess the articles by removing stopwords while retaining rare words, then upload the processed data back to the S3 bucket.

Question 15

A Generative AI engineering team is developing a supervised image recognition model to accurately identify pandas from a diverse dataset of wildlife photographs. Amazon SageMaker Ground Truth is used to label 1,000 panda images and split the data into training and testing sets, reserving 100 images as a constant test set. Amazon Rekognition Custom Labels is then configured to train a custom image classification model capable of recognizing distinctive panda features such as fur patterns, posture, and environmental background.  During the evaluation phase, the engineering team observes that the image classifier misclassifies several test images. A closer analysis reveals that in more than 75% of the misclassified images, the pandas appear upside down, indicating the model’s sensitivity to image orientation. The team must enhance the model’s ability to correctly identify pandas regardless of orientation, without collecting an entirely new dataset or extensively modifying the existing training pipeline.  Which approach will most effectively enhance the model’s accuracy in addressing this specific misclassification issue?

Accepted Answer

Expand the existing training dataset by introducing data augmentation techniques such as image rotation, flipping, and scaling.

Question 16

A global e-commerce company is developing a product return prediction model using the Amazon SageMaker XGBoost algorithm to identify which orders are most likely to result in customer returns. Transactional and behavioral data are stored in Amazon Redshift, and the data engineering team utilizes SageMaker Data Wrangler to clean and prepare training features, including product category, customer sentiment, and delivery delays.  After deploying the model, the team observes that while the training Area Under the Curve (AUC) is exceptionally high, the validation performance drops significantly. Model explainability reports from SageMaker Clarify reveal that the model is overly sensitive to a few dominant features like product price and region, suggesting overfitting to patterns in the training set. The team decides to tune the model’s hyperparameters to reduce model complexity and improve its ability to generalize to new data.  Which hyperparameter modification would best help reduce overfitting and improve validation performance for this XGBoost model?

Accepted Answer

Decrease the value of the max_depth hyperparameter to limit tree complexity and prevent overfitting on dominant patterns in the training data.

Question 17

A financial services company is developing a machine learning (ML) model in Amazon SageMaker AI to detect real-time fraudulent transactions. The company plans to integrate this model with Amazon Kinesis Data Streams for continuous transaction data ingestion and Amazon DynamoDB for storing transaction metadata.  The model must be trained on an extensive historical dataset stored in Amazon S3. The company must ensure that all data is encrypted in transit and at rest, and that access to sensitive data and model results is strictly controlled. The company also wants to track model performance over time to ensure that the model remains effective and that predictions align with business rules.  Which solution will secure the ML workflow, control access, and allow for continuous monitoring of model performance?

Accepted Answer

Use SageMaker AI to train the model with data from S3, enable S3 server-side encryption (SSE-KMS) for data at rest, configure IAM roles to control access to the model and data, and use Amazon CloudWatch to monitor model performance metrics and logging.

Question 18

A Generative AI engineering team at TD GenAILabs is developing a large-scale natural language processing system designed to process multilingual customer feedback from global data sources. The workflow uses Amazon SageMaker Data Wrangler to automate text cleaning, normalization, and tokenization, ensuring that all transformations are applied consistently before model training. The workflow also integrates Amazon Titan Text Embeddings to generate high-quality vector representations of text, enabling deeper semantic understanding and improving similarity analysis in later stages of the pipeline.  The team plans to train a Word2Vec-style model on SageMaker AI to generate embeddings that will support semantic similarity analysis and topic clustering. The dataset contains over one million sentences with inconsistent casing, mixed encodings, and minor typographical errors such as: “An Apple a DAY Keeps the doctor Away.” The preprocessing process must therefore ensure consistent sanitization, reproducibility, and embedding quality across multiple training cycles to maintain model accuracy and stability.  Which of the following operations should be implemented in the preprocessing phase to correctly sanitize and prepare the dataset for embedding and downstream predictions? (Select THREE.)

Accepted Answer

Segment the sentence into individual word units through tokenization., Exclude common non-informative words from the dataset using an English stop-word dictionary., Normalize the text by converting every word in the sentence to lowercase.

Question 19

A global insurance technology company built a claims automation system using Amazon SageMaker AI endpoint and Amazon Rekognition. Rekognition analyzes uploaded vehicle images to detect the severity and type of damage, while SageMaker uses this information, along with other structured features such as vehicle age, mileage, and estimated repair cost, to predict claim approval probability and estimated settlement time.  During initial training, the model assigned greater weight to the damage severity score and repair cost estimate as the most influential features for decision-making. However, recent observations suggest that the deployed model on the SageMaker AI endpoint might now be emphasizing vehicle age more heavily than damage severity, resulting in inconsistent predictions and potential policy compliance issues.  The data science team must implement a monitoring solution that detects when the feature weight or attribution importance of input variables shifts in production and automatically triggers alerts for investigation.  Which of the following should be implemented?

Accepted Answer

Use ModelExplainabilityMonitor class with a SHAP-based baseline to detect feature attribution drift in production. Regularly compare how the model assigns importance to input features against the baseline, and configure Amazon CloudWatch to alert stakeholders when attribution values drift beyond acceptable thresholds.

Question 20

A financial analytics enterprise is developing a secure generative AI platform that uses Amazon SageMaker AI for training, fine-tuning, and deploying proprietary foundation models, and Amazon Polly for generating natural-sounding speech from financial summaries and analytical insights. The platform is designed to unify both custom and managed AI capabilities, allowing internally developed models in SageMaker AI to work alongside Polly’s text-to-speech features for use cases such as voice-enabled financial dashboards, automated earnings calls, and interactive portfolio assistants while maintaining high levels of data privacy and regulatory compliance.  Due to strict data-governance requirements, the enterprise cloud security team has mandated that all AI workloads remain within a controlled network environment to eliminate direct internet access. The solution must ensure that SageMaker notebook instances operate securely within the private environment while maintaining compliance with organizational connectivity and security policies.  Which solution will enable this requirement?

Accepted Answer

Establish SageMaker VPC interface endpoints within the organization’s corporate VPC dedicated to AI operations.

Question 21

A data science team is running an anti-financial crime workload using Amazon SageMaker Training and SageMaker Feature Store. The team stores transactional features inside the Feature Store offline store, and those features are periodically retrieved and consumed by SageMaker model training jobs inside a scheduled retraining pipeline. The binary classifier is used for real-time fraud detection involving regulated payment flows, but the team is consistently observing a very high count of false negatives, even though the overall model accuracy is greater than 96%. The core issue is the severe class imbalance condition, where fraud transactions represent less than one-half of one percent of the entire dataset, and historical fraud samples remain extremely limited.  The goal is to increase the model’s ability to correctly detect fraudulent cases. The team wants to directly correct this imbalance problem before running the next retraining cycle.  Which solution will increase the fraudulent case detection performance?

Accepted Answer

Integrate a preprocessing step that applies the Synthetic Minority Oversampling Technique (SMOTE) on the minority fraudulent transaction class only before the training run begins.

Question 22

An AI development team at a publishing company is tasked with automating the summarization of large volumes of text, such as news articles, legal documents, and academic papers. The team needs to fine-tune a pre-trained large language model (LLM) for this task, but prefers a low-code/no-code (LCNC) solution. To ensure the summaries are accessible to a global audience, Amazon Translate is utilized to adapt the generated summaries for different regions and languages. The team now needs a solution to fine-tune the LLM with minimal manual intervention while integrating seamlessly with the existing workflow.  To meet these needs, Amazon SageMaker AI is being considered for fine-tuning the LLM, as it provides a low-code environment for experimentation. The team is focused on automating model training and fine-tuning to improve efficiency and reduce manual intervention, ensuring scalability and ease of use.  Which solution will best meet the team’s requirements?

Accepted Answer

Configure SageMaker Autopilot to fine-tune an LLM deployed via SageMaker JumpStart, streamlining model customization with automatic setup and minimal user intervention.

Question 23

A global e-commerce company is developing a product recommendation engine using Amazon SageMaker AI to personalize shopping experiences for millions of users. The data science team trains a deep learning model that predicts product affinity based on customer purchase history, browsing behavior, and geographic location. The model is deployed to a SageMaker real-time inference endpoint and integrated into the company’s recommendation API.  After the model goes live, the data ethics team discovers that specific product categories are being shown disproportionately to customers from certain regions. The organization must verify whether this imbalance is caused by underlying bias in the training dataset or by biased decision patterns in the deployed model predictions. The company must also generate automated reports that explain which input features most strongly influence the model’s recommendations to support transparency and compliance in personalization outcomes.  Which solution should the company implement to identify, measure, and explain potential bias across both the dataset and the model outputs?

Accepted Answer

Use SageMaker Clarify to detect and measure data bias, evaluate model fairness, and generate feature attribution explainability reports for compliance and transparency.

Question 24

A retail company is developing an AI-based recommendation system using Amazon Bedrock AgentCore. The system will leverage a large language model (LLM) fine-tuned in Amazon SageMaker AI to generate personalized product recommendations based on user browsing behavior.  The solution must ensure secure, real-time inference within the Bedrock environment, while also maintaining scalability and operational efficiency. Additionally, the company wants to monitor the inference latency and model performance metrics in real-time using Amazon CloudWatch.  Which solution ensures secure and scalable inference while minimizing operational overhead?

Accepted Answer

A retail company is developing an AI-based recommendation system using Amazon Bedrock AgentCore. The system will leverage a large language model (LLM) fine-tuned in Amazon SageMaker AI to generate personalized product recommendations based on user browsing behavior. The solution must ensure secure, real-time inference within the Bedrock environment, while also maintaining scalability and operational efficiency. Additionally, the company wants to monitor the inference latency and model performance metrics in real-time using Amazon CloudWatch. Which solution ensures secure and scalable inference while minimizing operational overhead?

Question 25

A large retail company is looking to leverage Amazon SageMaker AI notebooks and Amazon Comprehend to process and analyze customer feedback data. The company uses Amazon S3 buckets to store large datasets of customer reviews, and the data needs to be accessed securely for machine learning analysis.  Given the sensitivity of the data, the company mandates that all resources must remain within a secure Amazon Virtual Private Cloud (VPC). Additionally, all communication must occur over the AWS network.  Which solution will satisfy the given requirements?

Accepted Answer

Deploy the SageMaker AI notebook in a private subnet within a VPC and ensure that the VPC has private endpoints for both SageMaker AI and S3.

Question 26

A data science team at a retail company wants to predict customer churn based on various historical transactional data. The dataset contains 10,000 records, each with 1,500 attributes, such as demographic information, purchasing patterns, and customer service interactions. The team is looking for an automated way to build a model using Amazon SageMaker AI, which can predict whether a customer will churn while also identifying the most relevant features contributing to the prediction. To deepen insights, the team uses Amazon Comprehend to analyze sentiment and key phrases in customer interactions.  Which of the following solutions will best fulfill these requirements while minimizing manual effort?

Accepted Answer

Leverage SageMaker Autopilot to automatically train a classification model for forecasting customer churn. Then, utilize insights from SageMaker Clarify to determine which features most significantly influence the predictions.

Question 27

A data scientist is tasked with developing a fraud detection model for an e-commerce platform. The dataset consists of transaction records where fraudulent transactions are much less frequent than legitimate ones, resulting in a class imbalance.  To enrich the dataset, voice-based transaction logs or customer support calls may be transcribed using Amazon Transcribe and analyzed with Amazon Comprehend, but the core model will be built using structured data.  The data scientist has limited experience with advanced model development techniques but aims to minimize the operational overhead while ensuring the model is fair and unbiased. To accelerate development, the data scientist is planning to use Amazon SageMaker AI.  Which approach will fulfill the given requirements?

Accepted Answer

Use SageMaker Studio to preprocess and balance the data using the synthetic minority oversampling technique (SMOTE), then develop a fraud detection model with SageMaker JumpStart. Afterward, use SageMaker Clarify to check for bias and finalize the model for deployment.

Question 28

An e-commerce company leverages both Amazon SageMaker AI to train a machine learning (ML) model that predicts product-interest scores for users, and Amazon Personalize to generate real-time personalized product recommendations based on that model’s output.  The AI developer wants to visualize recommendation results across four distinct dimensions: the user’s interest score (first dimension) plotted against the product’s prior conversion rate (second dimension), then visualize a third dimension (product category) via color, and a fourth dimension (number of impressions for that product) via the size of each data point. The goal is to spot segments where high-interest, high-conversion products still receive low impressions.  Which approach best satisfies the given requirements?

Accepted Answer

Apply the SageMaker Canvas scatter plot visualization and map the third dimension (product category) to scatter point color and the fourth dimension (number of impressions) to scatter point size.

Question 29

A company operates a customer support chatbot that uses Amazon Bedrock to send user queries to an Amazon Nova large language model (LLM) for generating conversational responses. The chatbot is integrated with Amazon Kendra to retrieve relevant knowledge base articles, which are appended to the prompt before the request is sent to the model.  Users have reported that when similar questions are asked multiple times, different responses are sometimes returned, even though the retrieved Kendra results remain unchanged. The generative AI developer must configure the system to produce responses that are more consistent, deterministic, and less random, without modifying the knowledge retrieval process.  What approach solves these requirements?

Accepted Answer

Modify the inference parameters by lowering both the temperature value and the top_k sampling threshold.

Question 30

A company operates a fleet of wind turbines, solar panels, and weather stations in remote locations with unstable network connectivity. The company wants to improve its predictive maintenance capabilities by collecting and analyzing telemetry data from these devices, which include temperature, pressure, and rotational speed metrics. An AI developer plans to apply machine learning (ML) models, including anomaly detection and predictive maintenance algorithms, to identify potential failures in advance. The telemetry data will be stored securely in Amazon S3 for further analysis and integration with AWS machine learning services. The developer plans to use Amazon SageMaker AI for model training and deployment, while leveraging Amazon Comprehend to analyze unstructured logs that could contain important context about device performance. Which solution will meet the given requirements?

Accepted Answer

Route telemetry data over Message Queuing Telemetry Transport (MQTT) to AWS IoT Core, configure a rule in IoT Core to direct the data to an Amazon Data Firehose stream that delivers data to an S3.

Free AIP-C01 Practice Test Questions

Choose Your Practice Mode:

Download Real AIP-C01 Questions

Information

Top Exams

Hot Exams

About

Free AIP-C01 Practice Test Questions

Choose Your Practice Mode:

Flag Question for Review

Are you sure?

Download Real AIP-C01 Questions

Download Real AIP-C01 Questions

Information

Top Exams

Hot Exams

About

GDPR