AWS Certified Machine Learning Engineer – Associate practice questions

1. Question 1 of 25

   Case Study - A company is building a web-based AI application by using Amazon SageMaker. The application will provide the following capabilities and features: ML experimentation, training, a central model registry, model deployment, and model monitoring. The application must ensure secure and isolated use of training data during the ML lifecycle. The training data is stored in Amazon S3. The company needs to use the central model registry to manage different versions of models in the application. Which action will meet this requirement with the LEAST operational overhead?

   • ACreate a separate Amazon Elastic Container Registry (Amazon ECR) repository for each model.
   • BUse Amazon Elastic Container Registry (Amazon ECR) and unique tags for each model version.
   • CUse the SageMaker Model Registry and model groups to catalog the models.
   • DUse the SageMaker Model Registry and unique tags for each model version.

   Show answer

   Correct answer: C

   How others answered

   • C100%

   Explanation

   https://docs.aws.amazon.com/sagemaker/latest/dg/model-registry-models.html "Each model package in a Model Group corresponds to a trained model. The version of each model package is a numerical value that starts at 1 and is incremented with each new model package added to a Model Group. For example, if 5 model packages are added to a Model Group, the model package versions will be 1, 2, 3, 4, and 5."

   Adapted from community discussion by GiorgioGss · 2024-11-27 · 8 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

   Supporting references

   • AWS documentation
   • AWS documentation

   View community discussion4

   Candidate context

   Question 1 discussion

   These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

   1. G

      GiorgioGss

      2024-11-27

      Selected C8 upvotes

      https://docs.aws.amazon.com/sagemaker/latest/dg/model-registry-models.html "Each model package in a Model Group corresponds to a trained model. The version of each model package is a numerical value that starts at 1 and is incremented with each new model package added to a Model Group. For example, if 5 model packages are added to a Model Group, the model package versions will be 1, 2, 3, 4, and 5."

   2. K

      khchan123

      2025-01-01

      Selected C3 upvotes

      The correct answer is C. Use the SageMaker Model Registry and model groups to catalog the models. SageMaker Model Registry is specifically designed for managing ML models within the SageMaker ecosystem. It provides built-in versioning and cataloging capabilities for ML models. Model groups in the Model Registry allow for logical organization of related models. It integrates seamlessly with other SageMaker components like training, deployment, and monitoring. This solution offers the least operational overhead as it's a native SageMaker feature designed for this exact purpose. It provides features like model approval workflows, which are crucial for managing models in production environments.

   3. N

      Neo_2022

      2024-11-27

      Selected C4 upvotes

      https://aws.amazon.com/blogs/machine-learning/centralize-model-governance-with-sagemaker-model-registry-resource-access-manager-sharing/

   4. L

      Laxma99

      2025-04-02

      Selected C1 upvote

      SageMaker Model Registry is specifically designed for managing ML models within the SageMaker ecosystem. It provides built-in versioning and cataloging capabilities for ML models. Model groups in the Model Registry allow for logical organization of related models. It integrates seamlessly with other SageMaker components like training, deployment, and monitoring. This solution offers the least operational overhead as it's a native SageMaker feature designed for this exact purpose. It provides features like model approval workflows, which are crucial for managing models in production environments.

2. Question 2 of 25

   Case study - An ML engineer is developing a fraud detection model on AWS. The training dataset includes transaction logs, customer profiles, and tables from an on-premises MySQL database. The transaction logs and customer profiles are stored in Amazon S3. The dataset has a class imbalance that affects the learning of the model's algorithm. Additionally, many of the features have interdependencies. The algorithm is not capturing all the desired underlying patterns in the data. Which AWS service or feature can aggregate the data from the various data sources?

   • AAmazon EMR Spark jobs
   • BAmazon Kinesis Data Streams
   • CAmazon DynamoDB
   • DAWS Lake Formation

   Show answer

   Correct answer: D

   How others answered

   • D71%
   • A29%

   Only 71% agreed on this answer — treat it as contested rather than settled.

   Explanation

   Data Aggregation: Lake Formation is designed to create a data lake, a centralized repository that stores and manages data from various sources, including S3, relational databases (like MySQL), and other data sources. Data Transformation: It can transform and clean data, making it suitable for analysis and machine learning. This includes handling class imbalance and feature interdependencies. Data Access: It provides a unified interface to access data, simplifying the process of integrating data from different sources into the ML model. While other options like Amazon EMR Spark jobs and Amazon Kinesis Data Streams could be used for data processing and streaming, they are not the most efficient and straightforward solutions for this specific use case. Amazon DynamoDB is a NoSQL database, not designed for batch data processing and aggregation. Therefore, AWS Lake Formation is the best choice to aggregate and prepare the data for the ML model. ref:https://docs.aws.amazon.com/lake-formation/latest/dg/what-is-lake-formation.html

   Adapted from community discussion by TonyKean888 · 2024-12-10 · 4 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

   Supporting references

   • AWS documentation
   • AWS documentation

   View community discussion4

   Candidate context

   Question 2 discussion

   These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

   1. N

      ninomfr64

      2024-12-29

      Selected D5 upvotes

      Yet another poorly worded AWS certification question. Here is my reasoning, the question is about "aggregate the data from S3 and on-premise mysql" and I do intend "aggregate" as put in the same place, therefore: A. No, while EMR spark job can connect to S3 and MySQL (spark can connect to mysql database), but it is a better tool to process data and then sore them in S3 B. No, KDS it is for delivering streaming data sources to specific destinations (S3, OpenSearch ...) C. No, DynamoDB is a nosql db that is not a great fit here D. Yes, Lake Formation "combine different types of structured and unstructured data into a centralized repository" https://docs.aws.amazon.com/lake-formation/latest/dg/what-is-lake-formation.html and "with Lake Formation, you can import your data using workflows" and as it is based on AWS Glue it supports both S3 and mysql

   2. T

      TonyKean888

      2024-12-10

      Selected D4 upvotes

      Data Aggregation: Lake Formation is designed to create a data lake, a centralized repository that stores and manages data from various sources, including S3, relational databases (like MySQL), and other data sources. Data Transformation: It can transform and clean data, making it suitable for analysis and machine learning. This includes handling class imbalance and feature interdependencies. Data Access: It provides a unified interface to access data, simplifying the process of integrating data from different sources into the ML model. While other options like Amazon EMR Spark jobs and Amazon Kinesis Data Streams could be used for data processing and streaming, they are not the most efficient and straightforward solutions for this specific use case. Amazon DynamoDB is a NoSQL database, not designed for batch data processing and aggregation. Therefore, AWS Lake Formation is the best choice to aggregate and prepare the data for the ML model. ref:https://docs.aws.amazon.com/lake-formation/latest/dg/what-is-lake-formation.html

   3. A

      a4002bd

      2024-11-27

      Selected D7 upvotes

      Is it D? AWS Lake Formation ? EMR Spark jobs is more manual.

   4. D

      dbcert87

      2025-01-26

      Selected D5 upvotes

      AWS Lake Formation is a service designed to aggregate, catalog, and manage data from multiple data sources, including on-premises databases and Amazon S3, making it an ideal choice for this scenario. While Amazon EMR with Apache Spark is powerful for processing and analyzing large datasets, it focuses more on data processing than on data aggregation and cataloging. It doesn't inherently manage interdependencies or schema enforcement

3. Question 3 of 25

   A company shares Amazon SageMaker Studio notebooks that are accessible through a VPN. The company must enforce access controls to prevent malicious actors from exploiting presigned URLs to access the notebooks. Which solution will meet these requirements?

   • ASet up Studio client IP validation by using the aws:sourceIp IAM policy condition.
   • BSet up Studio client VPC validation by using the aws:sourceVpc IAM policy condition.
   • CSet up Studio client role endpoint validation by using the aws:PrimaryTag IAM policy condition.
   • DSet up Studio client user endpoint validation by using the aws:PrincipalTag IAM policy condition.

   Show answer

   Correct answer: A

   How others answered

   • A57%
   • B43%

   Only 57% agreed on this answer — treat it as contested rather than settled.

   Explanation

   ✅ Explanation: Context: The company is using Amazon SageMaker Studio notebooks. Access is allowed through a VPN, meaning users are coming from known, fixed IP ranges. The concern is unauthorized access via presigned URLs, which could potentially be used outside the trusted network. ✅ Why aws:sourceIp is the right choice: The aws:sourceIp condition in IAM policies allows you to restrict access based on the client's IP address. This is perfect for VPN-based setups where you know the IP range. It ensures that only users accessing from allowed IPs (e.g., your VPN subnet) can access SageMaker Studio resources, even if they have a valid presigned URL. This directly mitigates the risk of URL misuse from outside the VPN.

   Adapted from community discussion by eesa · 2025-03-21 · 2 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

   Supporting references

   • AWS documentation

   View community discussion3

   Candidate context

   Question 3 discussion

   These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

   1. E

      eesa

      2025-03-21

      Selected A2 upvotes

      ✅ Explanation: Context: The company is using Amazon SageMaker Studio notebooks. Access is allowed through a VPN, meaning users are coming from known, fixed IP ranges. The concern is unauthorized access via presigned URLs, which could potentially be used outside the trusted network. ✅ Why aws:sourceIp is the right choice: The aws:sourceIp condition in IAM policies allows you to restrict access based on the client's IP address. This is perfect for VPN-based setups where you know the IP range. It ensures that only users accessing from allowed IPs (e.g., your VPN subnet) can access SageMaker Studio resources, even if they have a valid presigned URL. This directly mitigates the risk of URL misuse from outside the VPN.

   2. C

      chris_spencer

      2025-03-12

      Selected A1 upvote

      A is correct. https://aws.amazon.com/blogs/machine-learning/secure-amazon-sagemaker-studio-presigned-urls-part-1-foundational-infrastructure/ Studio supports a few methods for enforcing access controls against presigned URL data exfiltration: Client IP validation using the IAM policy condition aws:sourceIp Client VPC validation using the IAM condition aws:sourceVpc Client VPC endpoint validation using the IAM policy condition aws:sourceVpce

   3. C

      chris_spencer

      2025-03-12

      1 upvote

      B is incorrect, IAM condition aws:sourceVpc is use for validating Client VPC. Client VPC validation using the IAM condition aws:sourceVpc

4. Question 4 of 25

   Case study - An ML engineer is developing a fraud detection model on AWS. The training dataset includes transaction logs, customer profiles, and tables from an on-premises MySQL database. The transaction logs and customer profiles are stored in Amazon S3. The dataset has a class imbalance that affects the learning of the model's algorithm. Additionally, many of the features have interdependencies. The algorithm is not capturing all the desired underlying patterns in the data. After the data is aggregated, the ML engineer must implement a solution to automatically detect anomalies in the data and to visualize the result. Which solution will meet these requirements?

   • AUse Amazon Athena to automatically detect the anomalies and to visualize the result.
   • BUse Amazon Redshift Spectrum to automatically detect the anomalies. Use Amazon QuickSight to visualize the result.
   • CUse Amazon SageMaker Data Wrangler to automatically detect the anomalies and to visualize the result.
   • DUse AWS Batch to automatically detect the anomalies. Use Amazon QuickSight to visualize the result.

   Show answer

   Correct answer: C

   How others answered

   • C100%

   Explanation

   SageMaker Data Wangler identify anomalies as part of the Data Quality and Insights Report (https://docs.aws.amazon.com/sagemaker/latest/dg/data-wrangler-data-insights.html) and provides various options for data visualization - https://docs.aws.amazon.com/sagemaker/latest/dg/data-wrangler-analyses.html

   Adapted from community discussion by ninomfr64 · 2024-12-29 · 2 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

   Supporting references

   • AWS documentation
   • AWS documentation

   View community discussion3

   Candidate context

   Question 4 discussion

   These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

   1. M

      motk123

      2024-12-09

      Selected C1 upvote

      Why Transform Categorical Data into Numerical Data? Machine learning algorithms generally require categorical data to be converted into numerical representations (e.g., one-hot encoding or embeddings) for training. Transforming numerical data into categorical data is unnecessary unless the problem explicitly requires it (e.g., binning for some specific applications). Why Use SageMaker Data Wrangler? Minimal Operational Overhead: Amazon SageMaker Data Wrangler provides a user-friendly interface to clean, preprocess, and transform data without needing to write custom code. Comprehensive Data Handling: Supports data sources like S3 and on-premises databases, and can handle both categorical and numerical data transformations efficiently. Why Not AWS Glue? AWS Glue is more suitable for large-scale ETL (Extract, Transform, Load) operations, such as schema inference or combining large datasets. It has higher operational overhead for specific ML data preprocessing tasks compared to SageMaker Data Wrangler.

   2. G

      GiorgioGss

      2024-11-27

      Selected C3 upvotes

      https://docs.aws.amazon.com/sagemaker/latest/dg/data-wrangler-analyses.html "Amazon SageMaker Data Wrangler includes built-in analyses that help you generate visualizations and data analyses in a few clicks. " This question is tricky because it makes you think you need Quicksight for the "visualization' part.

   3. N

      ninomfr64

      2024-12-29

      Selected C2 upvotes

      SageMaker Data Wangler identify anomalies as part of the Data Quality and Insights Report (https://docs.aws.amazon.com/sagemaker/latest/dg/data-wrangler-data-insights.html) and provides various options for data visualization - https://docs.aws.amazon.com/sagemaker/latest/dg/data-wrangler-analyses.html

5. Question 5 of 25

   Case study - An ML engineer is developing a fraud detection model on AWS. The training dataset includes transaction logs, customer profiles, and tables from an on-premises MySQL database. The transaction logs and customer profiles are stored in Amazon S3. The dataset has a class imbalance that affects the learning of the model's algorithm. Additionally, many of the features have interdependencies. The algorithm is not capturing all the desired underlying patterns in the data. The ML engineer needs to use an Amazon SageMaker built-in algorithm to train the model. Which algorithm should the ML engineer use to meet this requirement?

   • ALightGBM
   • BLinear learner
   • CК-means clustering
   • DNeural Topic Model (NTM)

   Show answer

   Correct answer: A

   How others answered

   • A57%
   • B43%

   Only 57% agreed on this answer — treat it as contested rather than settled.

   Explanation

   Here's why LightGBM is the most suitable algorithm for this fraud detection task: Handling Class Imbalance: LightGBM is particularly effective at handling imbalanced datasets, which is a key issue mentioned in the problem statement. It has built-in mechanisms to deal with class imbalance. Feature Interdependencies: LightGBM can capture complex feature interactions through its tree-based structure, addressing the issue of feature interdependencies mentioned in the problem. Capturing Underlying Patterns: As an advanced gradient boosting framework, LightGBM is excellent at capturing complex patterns in data, which the current algorithm is struggling with. Suitable for Fraud Detection: LightGBM is widely used in fraud detection tasks due to its high performance and ability to handle large datasets efficiently. Handling Various Data Types: It can work well with the mix of data types likely present in transaction logs, customer profiles, and database tables.

   Adapted from community discussion by khchan123 · 2025-01-01 · 3 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

   Supporting references

   • AWS documentation
   • AWS documentation

   View community discussion4

   Candidate context

   Question 5 discussion

   These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

   1. L

      Leo2023aws

      2024-11-28

      Selected A8 upvotes

      https://docs.aws.amazon.com/en_kr/sagemaker/latest/dg/lightgbm.html

   2. K

      khchan123

      2025-01-01

      Selected A3 upvotes

      Here's why LightGBM is the most suitable algorithm for this fraud detection task: Handling Class Imbalance: LightGBM is particularly effective at handling imbalanced datasets, which is a key issue mentioned in the problem statement. It has built-in mechanisms to deal with class imbalance. Feature Interdependencies: LightGBM can capture complex feature interactions through its tree-based structure, addressing the issue of feature interdependencies mentioned in the problem. Capturing Underlying Patterns: As an advanced gradient boosting framework, LightGBM is excellent at capturing complex patterns in data, which the current algorithm is struggling with. Suitable for Fraud Detection: LightGBM is widely used in fraud detection tasks due to its high performance and ability to handle large datasets efficiently. Handling Various Data Types: It can work well with the mix of data types likely present in transaction logs, customer profiles, and database tables.

   3. S

      Saransundar

      2024-12-05

      Selected A3 upvotes

      A. LightGBM: Handles class imbalance; captures feature interdependencies; models complex patterns. B. Linear Learner: Limited with interdependent features; struggles with complex patterns; suitable for linear relationships. C. K-means Clustering: Unsupervised algorithm; not suitable for classification; can't handle class imbalance. D. Neural Topic Model (NTM): Designed for topic modeling; unsuitable for fraud detection; doesn't address class imbalance.

   4. N

      ninomfr64

      2024-12-31

      Selected A2 upvotes

      We have an unbalanced dataset, this means we have labelled dataset thus we are going to use a supervised model training. This reduce options to A and B (K-means and NTM are unsupervised). Both LightGBM and Linear Learner provides hyperparameter to manage unbalanced datasets, respectively "scale-pos_weight" and "positive_example_weight_mult". I would go for LightGBM as this algorithm is more suited to handle complex relationship among features, while Linear Learner learns a linear function, or, for classification problems, a linear threshold function, and maps a vector x to an approximation of the label y.

6. Question 6 of 25

   A company has a team of data scientists who use Amazon SageMaker notebook instances to test ML models. When the data scientists need new permissions, the company attaches the permissions to each individual role that was created during the creation of the SageMaker notebook instance. The company needs to centralize management of the team's permissions. Which solution will meet this requirement?

   • ACreate a single IAM role that has the necessary permissions. Attach the role to each notebook instance that the team uses.
   • BCreate a single IAM group. Add the data scientists to the group. Associate the group with each notebook instance that the team uses.
   • CCreate a single IAM user. Attach the AdministratorAccess AWS managed IAM policy to the user. Configure each notebook instance to use the IAM user.
   • DCreate a single IAM group. Add the data scientists to the group. Create an IAM role. Attach the AdministratorAccess AWS managed IAM policy to the role. Associate the role with the group. Associate the group with each notebook instance that the team uses.

   Show answer

   Correct answer: A

   How others answered

   • A100%

   Explanation

   Creating a single IAM role allows centralized management of permissions for all SageMaker notebook instances. When permissions need to be updated, the changes are applied to the role, and all notebook instances automatically inherit the updated permissions. Why IAM Roles? IAM roles are the recommended way to provide permissions to AWS services like SageMaker because they securely delegate permissions without requiring long-term credentials. Why Not the Other Options? B. IAM groups manage permissions for users, not for AWS services or resources like SageMaker notebook instances. Groups cannot be attached directly to notebook instances. C. Using an IAM user with AdministratorAccess violates the principle of least privilege, granting unnecessary permissions. Additionally, IAM users are not intended to be attached to resources like notebook instances. D. This option combines unnecessary complexity (group and role association) and grants excessive permissions (AdministratorAccess), which is not secure or efficient.

   Adapted from community discussion by motk123 · 2024-12-10 · 3 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

   Supporting references

   • AWS documentation

   View community discussion3

   Candidate context

   Question 6 discussion

   These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

   1. M

      motk123

      2024-12-10

      Selected A3 upvotes

      Creating a single IAM role allows centralized management of permissions for all SageMaker notebook instances. When permissions need to be updated, the changes are applied to the role, and all notebook instances automatically inherit the updated permissions. Why IAM Roles? IAM roles are the recommended way to provide permissions to AWS services like SageMaker because they securely delegate permissions without requiring long-term credentials. Why Not the Other Options? B. IAM groups manage permissions for users, not for AWS services or resources like SageMaker notebook instances. Groups cannot be attached directly to notebook instances. C. Using an IAM user with AdministratorAccess violates the principle of least privilege, granting unnecessary permissions. Additionally, IAM users are not intended to be attached to resources like notebook instances. D. This option combines unnecessary complexity (group and role association) and grants excessive permissions (AdministratorAccess), which is not secure or efficient.

   2. N

      ninomfr64

      2024-12-31

      Selected A4 upvotes

      Yet another unclear question from AWS ... anyway, I am basically picking A as all the other options are not applicable or are unclear. A. Yes, this make sense B. No, you cannot assign (aka associate) group to notebook instances C. No, for two reason: AdministratorAccess policy is overly broad (violate least privilege principle) and you cannot assign IAM user to notebook instance D. No, for many reasons: AdministratorAccess policy is overly broad, not clear what associating a role to a group means (maybe a group has permissions to assume a role ...) and you cannot assign a group to a notebook

   3. G

      GiorgioGss

      2024-11-27

      Selected A3 upvotes

      Actually this is a best practice when working with notebooks in SageMaker. https://docs.aws.amazon.com/sagemaker/latest/dg/gs-setup-working-env.html

7. Question 7 of 25

   Case Study - A company is building a web-based AI application by using Amazon SageMaker. The application will provide the following capabilities and features: ML experimentation, training, a central model registry, model deployment, and model monitoring. The application must ensure secure and isolated use of training data during the ML lifecycle. The training data is stored in Amazon S3. The company is experimenting with consecutive training jobs. How can the company MINIMIZE infrastructure startup times for these jobs?

   • AUse Managed Spot Training.
   • BUse SageMaker managed warm pools.
   • CUse SageMaker Training Compiler.
   • DUse the SageMaker distributed data parallelism (SMDDP) library.

   Show answer

   Correct answer: B

   How others answered

   • B100%

   Explanation

   https://docs.aws.amazon.com/sagemaker/latest/dg/train-warm-pools.html#train-warm-pools-how-it-works SageMaker managed warm pools let you retain and reuse provisioned infrastructure after the completion of a training job to reduce latency for repetitive workloads, such as iterative experimentation or running many jobs consecutively.

   Adapted from community discussion by andy_10 · 2024-12-01 · 4 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

   Supporting references

   • AWS documentation
   • AWS documentation

   View community discussion4

   Candidate context

   Question 7 discussion

   These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

   1. A

      andy_10

      2024-12-01

      Selected B4 upvotes

      https://docs.aws.amazon.com/sagemaker/latest/dg/train-warm-pools.html#train-warm-pools-how-it-works SageMaker managed warm pools let you retain and reuse provisioned infrastructure after the completion of a training job to reduce latency for repetitive workloads, such as iterative experimentation or running many jobs consecutively.

   2. S

      S_201996

      2024-12-25

      Selected B3 upvotes

      SageMaker managed warm pools allow instances to stay in a ready state between consecutive training jobs, which minimizes infrastructure startup times. This feature is ideal for scenarios with frequent or consecutive training jobs, as it avoids the time-consuming process of provisioning infrastructure for each job.

   3. N

      ninomfr64

      2024-12-21

      Selected B1 upvote

      A. No, Managed Spot Training is used to reduce the compute cost for training B. Yes, Warm Pool allow to retain and re-use provisioned infrastructure, also use a persistent cache to store data across training job and help reduce infrastructure startup time as well as cost - https://docs.aws.amazon.com/sagemaker/latest/dg/train-warm-pools.html C. No, SageMaker Training Compiler is used to optimize your code for a specific target architecture D. No, the SageMaker distributed data parallelism (SMDDP) library is used parallelize training by distributing data across multiple instances. This doesn't reduce infrastructure startu ptime

   4. T

      tigrex73

      2024-11-27

      Selected B2 upvotes

      SageMaker managed warm pools are designed to reduce infrastructure startup times by keeping the training environment (instances, containers, and environment setup) ready between consecutive training jobs.

8. Question 8 of 25

   A company has a large, unstructured dataset. The dataset includes many duplicate records across several key attributes. Which solution on AWS will detect duplicates in the dataset with the LEAST code development?

   • AUse Amazon Mechanical Turk jobs to detect duplicates.
   • BUse Amazon QuickSight ML Insights to build a custom deduplication model.
   • CUse Amazon SageMaker Data Wrangler to pre-process and detect duplicates.
   • DUse the AWS Glue FindMatches transform to detect duplicates.

   Show answer

   Correct answer: D

   How others answered

   • D80%

   Explanation

   AWS Glue FindMatches is specifically designed to identify duplicate or matching records in datasets without requiring labeled training data. It uses machine learning to find fuzzy matches and allows customization to fine-tune the matching process, making it ideal for this scenario.

   Adapted from community discussion by Saransundar · 2024-12-05 · 7 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

   Supporting references

   • AWS documentation

   View community discussion3

   Candidate context

   Question 8 discussion

   These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

   1. S

      Saransundar

      2024-12-05

      Selected D7 upvotes

      AWS Glue FindMatches is specifically designed to identify duplicate or matching records in datasets without requiring labeled training data. It uses machine learning to find fuzzy matches and allows customization to fine-tune the matching process, making it ideal for this scenario.

   2. G

      GiorgioGss

      2024-11-27

      Selected D6 upvotes

      https://aws.amazon.com/about-aws/whats-new/2021/11/aws-glue-findmatches-new-data-existing-dataset/ "allows you to identify duplicate or matching records in your dataset"

   3. F

      feelgoodfactor

      2024-12-17

      Selected D3 upvotes

      The AWS Glue FindMatches transform is the most appropriate solution because it is specifically designed to detect duplicates, requires minimal development effort, and scales efficiently for large datasets.

9. Question 9 of 25

   An ML engineer has an Amazon Comprehend custom model in Account A in the us-east-1 Region. The ML engineer needs to copy the model to Account В in the same Region. Which solution will meet this requirement with the LEAST development effort?

   • AUse Amazon S3 to make a copy of the model. Transfer the copy to Account B.
   • BCreate a resource-based IAM policy. Use the Amazon Comprehend ImportModel API operation to copy the model to Account B.
   • CUse AWS DataSync to replicate the model from Account A to Account B.
   • DCreate an AWS Site-to-Site VPN connection between Account A and Account В to transfer the model.

   Show answer

   Correct answer: B

   How others answered

   • B100%

   Explanation

   Amazon Comprehend - ImportModel API to facilitate the transfer of custom models between AWS accounts. STEPS: 1. Exporting the model from Account A. 2. Creating a resource-based IAM policy in Account A to grant access to Account B. 3. Using the ImportModel API in Account B to import the model.

   Adapted from community discussion by Saransundar · 2024-12-05 · 3 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

   Supporting references

   • AWS documentation
   • AWS documentation

   View community discussion3

   Candidate context

   Question 9 discussion

   These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

   1. S

      Saransundar

      2024-12-05

      Selected B3 upvotes

      Amazon Comprehend - ImportModel API to facilitate the transfer of custom models between AWS accounts. STEPS: 1. Exporting the model from Account A. 2. Creating a resource-based IAM policy in Account A to grant access to Account B. 3. Using the ImportModel API in Account B to import the model.

   2. G

      GiorgioGss

      2024-11-27

      Selected B2 upvotes

      https://docs.aws.amazon.com/comprehend/latest/APIReference/API_ImportModel.html "The source model can be in your AWS account or another one."

   3. D

      dduenas

      2025-04-05

      Selected B1 upvote

      https://docs.aws.amazon.com/comprehend/latest/dg/custom-copy-importing.html

10. Question 10 of 25

    A company that has hundreds of data scientists is using Amazon SageMaker to create ML models. The models are in model groups in the SageMaker Model Registry. The data scientists are grouped into three categories: computer vision, natural language processing (NLP), and speech recognition. An ML engineer needs to implement a solution to organize the existing models into these groups to improve model discoverability at scale. The solution must not affect the integrity of the model artifacts and their existing groupings. Which solution will meet these requirements?

    • ACreate a custom tag for each of the three categories. Add the tags to the model packages in the SageMaker Model Registry.
    • BCreate a model group for each category. Move the existing models into these category model groups.
    • CUse SageMaker ML Lineage Tracking to automatically identify and tag which model groups should contain the models.
    • DCreate a Model Registry collection for each of the three categories. Move the existing model groups into the collections.

    Show answer

    Correct answer: D

    How others answered

    • D100%

    Explanation

    A could also be a valid option but in here we see exactly this: https://docs.aws.amazon.com/sagemaker/latest/dg/modelcollections.html "Any operation you perform on your Collections does not affect the integrity of the individual Model Groups they contain—the underlying Model Group artifacts in Amazon S3 and Amazon ECR are not modified."

    Adapted from community discussion by GiorgioGss · 2024-11-27 · 2 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion4

    Candidate context

    Question 10 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. G

       GiorgioGss

       2024-11-27

       Selected D2 upvotes

       A could also be a valid option but in here we see exactly this: https://docs.aws.amazon.com/sagemaker/latest/dg/modelcollections.html "Any operation you perform on your Collections does not affect the integrity of the individual Model Groups they contain—the underlying Model Group artifacts in Amazon S3 and Amazon ECR are not modified."

    2. L

       Linux_master

       2024-11-29

       Selected D3 upvotes

       https://docs.aws.amazon.com/sagemaker/latest/dg/modelcollections.html

    3. A

       a4002bd

       2024-11-27

       Selected D2 upvotes

       I pick D. Creating custom tags for each of the three categories and adding them to the model packages in the SageMaker Model Registry (Option A) is a valid approach. However, it might not be as effective for organizing models at scale compared to using Model Registry collections.

    4. A

       abrarjahin

       2025-01-21

       Selected D2 upvotes

       Because according to the documentation - "Any operation you perform on your Collections does not affect the integrity of the individual Model Groups they contain—the underlying Model Group artifacts in Amazon S3 and Amazon ECR are not modified."

11. Question 11 of 25

    A company is using ML to predict the presence of a specific weed in a farmer's field. The company is using the Amazon SageMaker linear learner built-in algorithm with a value of multiclass_dassifier for the predictorjype hyperparameter. What should the company do to MINIMIZE false positives?

    • ASet the value of the weight decay hyperparameter to zero.
    • BIncrease the number of training epochs.
    • CIncrease the value of the target_precision hyperparameter.
    • DChange the value of the predictorjype hyperparameter to regressor.

    Show answer

    Correct answer: C

    How others answered

    • C100%

    Explanation

    A. Weight decay = 0 → No regularization, doesn’t target false positives. B. More epochs → Longer training, risks overfitting, no direct impact on false positives. C. Higher precision → Prioritizes correct positives, reduces false positives. D. Regressor → Predicts continuous values, unsuitable for classification.

    Adapted from community discussion by Saransundar · 2024-12-04 · 2 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion2

    Candidate context

    Question 11 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. S

       Saransundar

       2024-12-04

       Selected C2 upvotes

       A. Weight decay = 0 → No regularization, doesn’t target false positives. B. More epochs → Longer training, risks overfitting, no direct impact on false positives. C. Higher precision → Prioritizes correct positives, reduces false positives. D. Regressor → Predicts continuous values, unsuitable for classification.

    2. G

       GiorgioGss

       2024-11-27

       Selected C1 upvote

       https://docs.aws.amazon.com/sagemaker/latest/dg/ll_hyperparameters.html

12. Question 12 of 25

    A company has implemented a data ingestion pipeline for sales transactions from its ecommerce website. The company uses Amazon Data Firehose to ingest data into Amazon OpenSearch Service. The buffer interval of the Firehose stream is set for 60 seconds. An OpenSearch linear model generates real-time sales forecasts based on the data and presents the data in an OpenSearch dashboard. The company needs to optimize the data ingestion pipeline to support sub-second latency for the real-time dashboard. Which change to the architecture will meet these requirements?

    • AUse zero buffering in the Firehose stream. Tune the batch size that is used in the PutRecordBatch operation.
    • BReplace the Firehose stream with an AWS DataSync task. Configure the task with enhanced fan-out consumers.
    • CIncrease the buffer interval of the Firehose stream from 60 seconds to 120 seconds.
    • DReplace the Firehose stream with an Amazon Simple Queue Service (Amazon SQS) queue.

    Show answer

    Correct answer: A

    How others answered

    • A100%

    Explanation

    A. Use zero buffering to minimize latency by delivering data immediate. Tune batch size to optimize throughput & ensures sub-second delivery for real-time dashboards.

    Adapted from community discussion by Saransundar · 2024-12-05 · 3 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion2

    Candidate context

    Question 12 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. S

       Saransundar

       2024-12-05

       Selected A3 upvotes

       A. Use zero buffering to minimize latency by delivering data immediate. Tune batch size to optimize throughput & ensures sub-second delivery for real-time dashboards.

    2. G

       GiorgioGss

       2024-11-27

       Selected A3 upvotes

       Although is quite new solution , A will do the trick: https://aws.amazon.com/about-aws/whats-new/2023/12/amazon-kinesis-data-firehose-zero-buffering/

13. Question 13 of 25

    An ML engineer needs to use an Amazon EMR cluster to process large volumes of data in batches. Any data loss is unacceptable. Which instance purchasing option will meet these requirements MOST cost-effectively?

    • ARun the primary node, core nodes, and task nodes on On-Demand Instances.
    • BRun the primary node, core nodes, and task nodes on Spot Instances.
    • CRun the primary node on an On-Demand Instance. Run the core nodes and task nodes on Spot Instances.
    • DRun the primary node and core nodes on On-Demand Instances. Run the task nodes on Spot Instances.

    Show answer

    Correct answer: D

    How others answered

    • D100%

    Explanation

    https://docs.aws.amazon.com/emr/latest/ManagementGuide/emr-plan-instances-guidelines.html#emr-plan-spot-instances "The task nodes process data but do not hold persistent data in HDFS. If they terminate because the Spot price has risen above your maximum Spot price, no data is lost"

    Adapted from community discussion by GiorgioGss · 2024-11-27 · 4 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion3

    Candidate context

    Question 13 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. G

       GiorgioGss

       2024-11-27

       Selected D4 upvotes

       https://docs.aws.amazon.com/emr/latest/ManagementGuide/emr-plan-instances-guidelines.html#emr-plan-spot-instances "The task nodes process data but do not hold persistent data in HDFS. If they terminate because the Spot price has risen above your maximum Spot price, no data is lost"

    2. S

       Saransundar

       2024-12-05

       Selected D1 upvote

       Acceptable data loss: Spot can be used but you can't change an instance purchasing option while a cluster is running. To change from On-Demand to Spot Instances or vice versa, for the primary and core nodes, you must terminate the cluster and launch a new one. For task nodes, you can launch a new task instance group or instance fleet, and remove the old one.

    3. S

       Saransundar

       2024-12-05

       1 upvote

       Unacceptable data loss: Primary node : Cluster doesn't start sometimes If spot only used Core nodes: Possible of partial data loss HDFS Task nodes: No data loss and do not hold persistent data

14. Question 14 of 25

    Case Study - A company is building a web-based AI application by using Amazon SageMaker. The application will provide the following capabilities and features: ML experimentation, training, a central model registry, model deployment, and model monitoring. The application must ensure secure and isolated use of training data during the ML lifecycle. The training data is stored in Amazon S3. The company needs to run an on-demand workflow to monitor bias drift for models that are deployed to real-time endpoints from the application. Which action will meet this requirement?

    • AConfigure the application to invoke an AWS Lambda function that runs a SageMaker Clarify job.
    • BInvoke an AWS Lambda function to pull the sagemaker-model-monitor-analyzer built-in SageMaker image.
    • CUse AWS Glue Data Quality to monitor bias.
    • DUse SageMaker notebooks to compare the bias.

    Show answer

    Correct answer: A

    How others answered

    • A100%

    Explanation

    A. Yes, Clarify allows to get bias - https://docs.aws.amazon.com/sagemaker/latest/dg/clarify-configure-processing-jobs.html B. No, the built-in image sagemaker-model-monitor-analyzer provides a range of model monitoring capabilities (constraint suggestion, statistics generation, constraint validation against a baseline, and emitting Amazon CloudWatch metrics) but you need Clarify for bias C. No, Glue Data Quality doesn't analyze bias D. No, well from a Notebook you can execute pretty much everything including a Clarify Job, however notebooks are for experiments and models development not for enabling real-time application features

    Adapted from community discussion by ninomfr64 · 2024-12-21 · 6 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation
    • AWS documentation

    View community discussion4

    Candidate context

    Question 14 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. N

       ninomfr64

       2024-12-21

       Selected A6 upvotes

       A. Yes, Clarify allows to get bias - https://docs.aws.amazon.com/sagemaker/latest/dg/clarify-configure-processing-jobs.html B. No, the built-in image sagemaker-model-monitor-analyzer provides a range of model monitoring capabilities (constraint suggestion, statistics generation, constraint validation against a baseline, and emitting Amazon CloudWatch metrics) but you need Clarify for bias C. No, Glue Data Quality doesn't analyze bias D. No, well from a Notebook you can execute pretty much everything including a Clarify Job, however notebooks are for experiments and models development not for enabling real-time application features

    2. T

       tigrex73

       2024-11-27

       Selected A2 upvotes

       SageMaker Clarify is a tool designed to detect and monitor bias in datasets and models. It provides built-in capabilities for bias analysis, both pre-training (data bias) and post-training (model bias). Using AWS Lambda to invoke the job ensures automation and on-demand execution, reducing operational complexity while meeting the requirement for monitoring bias drift.

    3. G

       GiorgioGss

       2024-11-27

       Selected A2 upvotes

       https://docs.aws.amazon.com/sagemaker/latest/dg/clarify-measure-data-bias.html

    4. L

       Laxma99

       2025-04-02

       Selected A1 upvote

       SageMaker Clarify can be used to analyze bias drift in models. By integrating this with a Lambda function, the workflow can be triggered on-demand whenever the application requires bias monitoring.

15. Question 15 of 25

    An ML engineer needs to use AWS CloudFormation to create an ML model that an Amazon SageMaker endpoint will host. Which resource should the ML engineer declare in the CloudFormation template to meet this requirement?

    • AAWS::SageMaker::Model
    • BAWS::SageMaker::Endpoint
    • CAWS::SageMaker::NotebookInstance
    • DAWS::SageMaker::Pipeline

    Show answer

    Correct answer: A

    How others answered

    • A100%

    Explanation

    Type: AWS::SageMaker::Model Properties: Containers: - ContainerDefinition EnableNetworkIsolation: Boolean ExecutionRoleArn: String InferenceExecutionConfig: InferenceExecutionConfig ModelName: String PrimaryContainer: ContainerDefinition Tags: - Tag VpcConfig: VpcConfig

    Adapted from community discussion by Saransundar · 2024-12-05 · 2 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion3

    Candidate context

    Question 15 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. S

       Saransundar

       2024-12-05

       Selected A2 upvotes

       Type: AWS::SageMaker::Model Properties: Containers: - ContainerDefinition EnableNetworkIsolation: Boolean ExecutionRoleArn: String InferenceExecutionConfig: InferenceExecutionConfig ModelName: String PrimaryContainer: ContainerDefinition Tags: - Tag VpcConfig: VpcConfig

    2. E

       eesa

       2025-03-21

       Selected A1 upvote

       A. AWS::SageMaker::Model ✅ This resource defines the model, including: Location of model artifacts in S3 Inference image/container IAM role It is a prerequisite for deploying a model to an endpoint.

    3. G

       GiorgioGss

       2024-11-27

       Selected A1 upvote

       https://docs.aws.amazon.com/AWSCloudFormation/latest/UserGuide/aws-resource-sagemaker-model.html

16. Question 16 of 25

    An ML engineer is evaluating several ML models and must choose one model to use in production. The cost of false negative predictions by the models is much higher than the cost of false positive predictions. Which metric finding should the ML engineer prioritize the MOST when choosing the model?

    • ALow precision
    • BHigh precision
    • CLow recall
    • DHigh recall

    Show answer

    Correct answer: D

    How others answered

    • D100%

    Explanation

    A. Low precision: Increases false positives; less relevant here. B. High precision: Reduces false positives; not the priority. C. Low recall: Increases false negatives; must be avoided. D. High recall: Correct; minimizes false negatives.

    Adapted from community discussion by Saransundar · 2024-12-05 · 2 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion2

    Candidate context

    Question 16 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. S

       Saransundar

       2024-12-05

       Selected D2 upvotes

       A. Low precision: Increases false positives; less relevant here. B. High precision: Reduces false positives; not the priority. C. Low recall: Increases false negatives; must be avoided. D. High recall: Correct; minimizes false negatives.

    2. G

       GiorgioGss

       2024-11-27

       Selected D1 upvote

       https://docs.aws.amazon.com/comprehend/latest/dg/cer-doc-class.html "High recall means that the classifier returned most of the relevant results."

17. Question 17 of 25

    A credit card company has a fraud detection model in production on an Amazon SageMaker endpoint. The company develops a new version of the model. The company needs to assess the new model's performance by using live data and without affecting production end users. Which solution will meet these requirements?

    • ASet up SageMaker Debugger and create a custom rule.
    • BSet up blue/green deployments with all-at-once traffic shifting.
    • CSet up blue/green deployments with canary traffic shifting.
    • DSet up shadow testing with a shadow variant of the new model.

    Show answer

    Correct answer: D

    How others answered

    • D100%

    Explanation

    Shadow testing is a technique used to evaluate a new model's performance by running it alongside the current production model, processing the same live data but without affecting production outcomes.

    Adapted from community discussion by Saransundar · 2024-12-05 · 2 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion2

    Candidate context

    Question 17 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. S

       Saransundar

       2024-12-05

       Selected D2 upvotes

       Shadow testing is a technique used to evaluate a new model's performance by running it alongside the current production model, processing the same live data but without affecting production outcomes.

    2. G

       GiorgioGss

       2024-11-27

       Selected D1 upvote

       https://docs.aws.amazon.com/sagemaker/latest/dg/shadow-tests-create.html

18. Question 18 of 25

    An ML engineer trained an ML model on Amazon SageMaker to detect automobile accidents from dosed-circuit TV footage. The ML engineer used SageMaker Data Wrangler to create a training dataset of images of accidents and non-accidents. The model performed well during training and validation. However, the model is underperforming in production because of variations in the quality of the images from various cameras. Which solution will improve the model's accuracy in the LEAST amount of time?

    • ACollect more images from all the cameras. Use Data Wrangler to prepare a new training dataset.
    • BRecreate the training dataset by using the Data Wrangler corrupt image transform. Specify the impulse noise option.
    • CRecreate the training dataset by using the Data Wrangler enhance image contrast transform. Specify the Gamma contrast option.
    • DRecreate the training dataset by using the Data Wrangler resize image transform. Crop all images to the same size.

    Show answer

    Correct answer: B

    How others answered

    • B75%
    • C25%

    Explanation

    It's B. Did you guys clearly understand the question? "The model performed well during training and validation. However, the model is underperforming in production because of variations in the quality of the images from various cameras." https://aws.amazon.com/blogs/machine-learning/prepare-image-data-with-amazon-sagemaker-data-wrangler/ Corrupting an image or creating any kind of noise helps make a model more robust. The model can predict with more accuracy even if it receives a corrupted image because it was trained with corrupt and non-corrupt images.

    Adapted from community discussion by Lance665 · 2024-12-21 · 6 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion3

    Candidate context

    Question 18 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. L

       Lance665

       2024-12-21

       Selected B6 upvotes

       It's B. Did you guys clearly understand the question? "The model performed well during training and validation. However, the model is underperforming in production because of variations in the quality of the images from various cameras." https://aws.amazon.com/blogs/machine-learning/prepare-image-data-with-amazon-sagemaker-data-wrangler/ Corrupting an image or creating any kind of noise helps make a model more robust. The model can predict with more accuracy even if it receives a corrupted image because it was trained with corrupt and non-corrupt images.

    2. 6

       6913a18

       2025-01-03

       Selected B2 upvotes

       Corrupting an image or creating any kind of noise helps make a model more robust. The model can predict with more accuracy even if it receives a corrupted image because it was trained with corrupt and non-corrupt images. https://aws.amazon.com/blogs/machine-learning/prepare-image-data-with-amazon-sagemaker-data-wrangler/

    3. E

       eesa

       2025-03-31

       Selected B1 upvote

       Recreate the training dataset by using the Data Wrangler corrupt image transform. Specify the impulse noise option. ✅ Adds noise to images, simulating low-quality or distorted camera feeds. ✅ Fast — can be done with existing data. ✅ Makes model more robust to variations in image quality. ✅ Best for improving accuracy in the least amount of time.

19. Question 19 of 25

    A company is building a deep learning model on Amazon SageMaker. The company uses a large amount of data as the training dataset. The company needs to optimize the model's hyperparameters to minimize the loss function on the validation dataset. Which hyperparameter tuning strategy will accomplish this goal with the LEAST computation time?

    • AHyperband
    • BGrid search
    • CBayesian optimization
    • DRandom search

    Show answer

    Correct answer: A

    How others answered

    • A100%

    Explanation

    A. Hyperband: Efficient & best --> Right answer B. Grid Search: Exhaustive and tries all combos C. Bayesian Optimization: Smart with best combination D. Random Search: Random

    Adapted from community discussion by Saransundar · 2024-12-04 · 2 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion2

    Candidate context

    Question 19 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. S

       Saransundar

       2024-12-04

       Selected A2 upvotes

       A. Hyperband: Efficient & best --> Right answer B. Grid Search: Exhaustive and tries all combos C. Bayesian Optimization: Smart with best combination D. Random Search: Random

    2. G

       GiorgioGss

       2024-11-28

       Selected A1 upvote

       https://aws.amazon.com/blogs/machine-learning/amazon-sagemaker-automatic-model-tuning-now-provides-up-to-three-times-faster-hyperparameter-tuning-with-hyperband/ "efficient resource utilization and a better time-to-convergence."

20. Question 20 of 25

    A company needs to host a custom ML model to perform forecast analysis. The forecast analysis will occur with predictable and sustained load during the same 2-hour period every day. Multiple invocations during the analysis period will require quick responses. The company needs AWS to manage the underlying infrastructure and any auto scaling activities. Which solution will meet these requirements?

    • ASchedule an Amazon SageMaker batch transform job by using AWS Lambda.
    • BConfigure an Auto Scaling group of Amazon EC2 instances to use scheduled scaling.
    • CUse Amazon SageMaker Serverless Inference with provisioned concurrency.
    • DRun the model on an Amazon Elastic Kubernetes Service (Amazon EKS) cluster on Amazon EC2 with pod auto scaling.

    Show answer

    Correct answer: C

    How others answered

    • C100%

    Explanation

    Load is predictable and sustainable with 2 hrs usage pattern; Needs quick response as well; Sagemaker - Provisioned concurrency + Serverless inference will be able to support it. https://docs.aws.amazon.com/sagemaker/latest/dg/serverless-endpoints.html

    Adapted from community discussion by Saransundar · 2024-12-04 · 2 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion2

    Candidate context

    Question 20 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. S

       Saransundar

       2024-12-04

       Selected C2 upvotes

       Load is predictable and sustainable with 2 hrs usage pattern; Needs quick response as well; Sagemaker - Provisioned concurrency + Serverless inference will be able to support it. https://docs.aws.amazon.com/sagemaker/latest/dg/serverless-endpoints.html

    2. G

       GiorgioGss

       2024-11-28

       Selected C2 upvotes

       For only 2h each day it does not worth provisioning resources. Go serverless.

21. Question 21 of 25

    A company wants to develop an ML model by using tabular data from its customers. The data contains meaningful ordered features with sensitive information that should not be discarded. An ML engineer must ensure that the sensitive data is masked before another team starts to build the model. Which solution will meet these requirements?

    • AUse Amazon Made to categorize the sensitive data.
    • BPrepare the data by using AWS Glue DataBrew.
    • CRun an AWS Batch job to change the sensitive data to random values.
    • DRun an Amazon EMR job to change the sensitive data to random values.

    Show answer

    Correct answer: B

    How others answered

    • B77%
    • A23%

    Explanation

    AWS Glue DataBrew (Option B) is the most efficient and user-friendly solution for masking sensitive information while retaining the structure and order of tabular data, making it ideal for preparing data for ML model development. AWS Macie cannot mask data.

    Adapted from community discussion by ZoroRoronoa · 2024-12-28 · 4 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion4

    Candidate context

    Question 21 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. Z

       ZoroRoronoa

       2024-12-28

       Selected B4 upvotes

       AWS Glue DataBrew (Option B) is the most efficient and user-friendly solution for masking sensitive information while retaining the structure and order of tabular data, making it ideal for preparing data for ML model development. AWS Macie cannot mask data.

    2. G

       GiorgioGss

       2024-11-28

       Selected B3 upvotes

       https://docs.aws.amazon.com/databrew/latest/dg/personal-information-protection.html

    3. E

       eesa

       2025-03-31

       Selected B1 upvote

       Prepare the data by using AWS Glue DataBrew ✅ Purpose-built for data preparation — works with tabular data. ✅ Offers data masking, pseudonymization, and data transformations. ✅ No-code/low-code tool suitable for collaborative workflows between engineers and analysts. ✅ Keeps the order and structure of features intact. ✅ Best fit for masking sensitive info before ML modeling.

    4. S

       Saransundar

       2024-12-05

       Selected B2 upvotes

       Customer tabular data → AWS Glue DataBrew → Mask sensitive data → Prepare for ML model building

22. Question 22 of 25

    A company has a binary classification model in production. An ML engineer needs to develop a new version of the model. The new model version must maximize correct predictions of positive labels and negative labels. The ML engineer must use a metric to recalibrate the model to meet these requirements. Which metric should the ML engineer use for the model recalibration?

    • AAccuracy
    • BPrecision
    • CRecall
    • DSpecificity

    Show answer

    Correct answer: A

    How others answered

    • A100%

    Explanation

    A. Accuracy: Correct choice; maximizes both true positives and true negatives. Formula: (TP + TN) / Total Predictions B. Precision: Focuses only on true positives, not negatives. Formula: TP / (TP + FP) C. Recall: Focuses on capturing all true positives, ignoring negatives. Formula: TP / (TP + FN) D. Specificity: Focuses only on true negatives, ignoring positives. Formula: TN / (TN + FP)

    Adapted from community discussion by Saransundar · 2024-12-05 · 3 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion4

    Candidate context

    Question 22 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. S

       Saransundar

       2024-12-05

       Selected A3 upvotes

       A. Accuracy: Correct choice; maximizes both true positives and true negatives. Formula: (TP + TN) / Total Predictions B. Precision: Focuses only on true positives, not negatives. Formula: TP / (TP + FP) C. Recall: Focuses on capturing all true positives, ignoring negatives. Formula: TP / (TP + FN) D. Specificity: Focuses only on true negatives, ignoring positives. Formula: TN / (TN + FP)

    2. A

       aws_Tamilan

       2025-03-31

       Selected A1 upvote

       🔑 Keyword: Maximize correct predictions of both positive and negative labels ✅ Correct Answer: A. Accuracy Why? Accuracy measures the proportion of correctly classified instances (both positive and negative). It is the most appropriate metric when both false positives and false negatives need to be minimized. Why Others Are Wrong? ❌ B. Precision focuses only on correctly predicted positives, not overall correctness. ❌ C. Recall focuses only on capturing all true positives, ignoring true negatives. ❌ D. Specificity only measures the ability to identify true negatives.

    3. K

       kyo

       2025-01-19

       Selected A1 upvote

       For imbalanced data, F1 score is preferred over confusion matrix-derived metrics. Ideally, option A would be "F1 Score". Given the choices, A (Accuracy) is the most reasonable, though not ideal.

    4. G

       GiorgioGss

       2024-11-28

       Selected A1 upvote

       https://docs.aws.amazon.com/sagemaker/latest/dg/clarify-accuracy-evaluation.html

23. Question 23 of 25

    A company is using Amazon SageMaker to create ML models. The company's data scientists need fine-grained control of the ML workflows that they orchestrate. The data scientists also need the ability to visualize SageMaker jobs and workflows as a directed acyclic graph (DAG). The data scientists must keep a running history of model discovery experiments and must establish model governance for auditing and compliance verifications. Which solution will meet these requirements?

    • AUse AWS CodePipeline and its integration with SageMaker Studio to manage the entire ML workflows. Use SageMaker ML Lineage Tracking for the running history of experiments and for auditing and compliance verifications.
    • BUse AWS CodePipeline and its integration with SageMaker Experiments to manage the entire ML workflows. Use SageMaker Experiments for the running history of experiments and for auditing and compliance verifications.
    • CUse SageMaker Pipelines and its integration with SageMaker Studio to manage the entire ML workflows. Use SageMaker ML Lineage Tracking for the running history of experiments and for auditing and compliance verifications.
    • DUse SageMaker Pipelines and its integration with SageMaker Experiments to manage the entire ML workflows. Use SageMaker Experiments for the running history of experiments and for auditing and compliance verifications.

    Show answer

    Correct answer: C

    How others answered

    • C67%
    • D33%

    Only 67% agreed on this answer — treat it as contested rather than settled.

    Explanation

    The correct answer is C. Options B and D suggest using SageMaker Experiments, which is good for tracking experiments but doesn't provide the comprehensive lineage tracking and governance features that ML Lineage Tracking offers. Running history of experiments: SageMaker ML Lineage Tracking provides a comprehensive way to track the lineage of ML workflows, including datasets, algorithms, hyperparameters, and models. This fulfills the requirement for keeping a running history of model discovery experiments. Model governance for auditing and compliance: ML Lineage Tracking also supports model governance by providing detailed information about each step in the ML process, which is crucial for auditing and compliance verifications.

    Adapted from community discussion by khchan123 · 2025-01-03 · 4 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation
    • AWS documentation

    View community discussion4

    Candidate context

    Question 23 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. K

       khchan123

       2025-01-03

       Selected C4 upvotes

       The correct answer is C. Options B and D suggest using SageMaker Experiments, which is good for tracking experiments but doesn't provide the comprehensive lineage tracking and governance features that ML Lineage Tracking offers. Running history of experiments: SageMaker ML Lineage Tracking provides a comprehensive way to track the lineage of ML workflows, including datasets, algorithms, hyperparameters, and models. This fulfills the requirement for keeping a running history of model discovery experiments. Model governance for auditing and compliance: ML Lineage Tracking also supports model governance by providing detailed information about each step in the ML process, which is crucial for auditing and compliance verifications.

    2. G

       GiorgioGss

       2024-11-28

       Selected C5 upvotes

       https://docs.aws.amazon.com/sagemaker/latest/dg/define-pipeline.html I don't see how you can manage the "entire ML flow" (as question asks) with something else other than Studio.

    3. S

       Saransundar

       2024-12-04

       Selected C4 upvotes

       https://docs.aws.amazon.com/sagemaker/latest/dg/lineage-tracking.html With SageMaker AI Lineage Tracking data scientists and model builders to Keep a running history of model discovery experiments. Establish model governance by tracking model lineage artifacts for auditing and compliance verification.

    4. A

       aws_Tamilan

       2025-03-31

       Selected C1 upvote

       🔑 Keyword: Fine-grained control, DAG visualization, model governance, auditing ✅ Correct Answer: C. Use SageMaker Pipelines and its integration with SageMaker Studio to manage the entire ML workflows. Use SageMaker ML Lineage Tracking for the running history of experiments and for auditing and compliance verifications. Why? SageMaker Pipelines provides workflow orchestration with DAG visualization. SageMaker ML Lineage Tracking enables experiment history and model governance for auditing. Why Others Are Wrong? ❌ A & B. AWS CodePipeline is more suited for software CI/CD than for ML workflows. ❌ D. SageMaker Experiments focuses on tracking experiments but lacks full model lineage tracking.

24. Question 24 of 25

    A company needs to create a central catalog for all the company's ML models. The models are in AWS accounts where the company developed the models initially. The models are hosted in Amazon Elastic Container Registry (Amazon ECR) repositories. Which solution will meet these requirements?

    • AConfigure ECR cross-account replication for each existing ECR repository. Ensure that each model is visible in each AWS account.
    • BCreate a new AWS account with a new ECR repository as the central catalog. Configure ECR cross-account replication between the initial ECR repositories and the central catalog.
    • CUse the Amazon SageMaker Model Registry to create a model group for models hosted in Amazon ECR. Create a new AWS account. In the new account, use the SageMaker Model Registry as the central catalog. Attach a cross-account resource policy to each model group in the initial AWS accounts.
    • DUse an AWS Glue Data Catalog to store the models. Run an AWS Glue crawler to migrate the models from the ECR repositories to the Data Catalog. Configure cross-account access to the Data Catalog.

    Show answer

    Correct answer: C

    How others answered

    • C88%

    Explanation

    The question asks for a "central catalog" so I believe metadata, lineage tracking are also "included". ECR could not be the solution.

    Adapted from community discussion by GiorgioGss · 2024-11-28 · 6 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion4

    Candidate context

    Question 24 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. G

       GiorgioGss

       2024-11-28

       Selected C6 upvotes

       The question asks for a "central catalog" so I believe metadata, lineage tracking are also "included". ECR could not be the solution.

    2. A

       aws_Tamilan

       2025-03-31

       Selected C1 upvote

       🔑 Keyword: Central catalog for ML models across AWS accounts, models stored in ECR ✅ Correct Answer: C. Use the Amazon SageMaker Model Registry to create a model group for models hosted in Amazon ECR. Create a new AWS account. In the new account, use the SageMaker Model Registry as the central catalog. Attach a cross-account resource policy to each model group in the initial AWS accounts. Why? SageMaker Model Registry provides a central catalog for ML models, supporting cross-account access. Cross-account resource policies enable access to models across AWS accounts. Why Others Are Wrong? ❌ A. ECR cross-account replication does not create a centralized catalog. ❌ B. A new ECR repository with replication does not offer model versioning or governance. ❌ D. AWS Glue Data Catalog is for metadata storage, not for ML model management.

    3. E

       Ell89

       2024-12-31

       1 upvote

       yes. https://aws.amazon.com/about-aws/whats-new/2023/10/amazon-sagemaker-model-registry-private-model-repositories/

    4. S

       Saransundar

       2024-12-04

       1 upvote

       Is it possible to use SageMaker Model Registry to create a model group for models from ECR repos?

25. Question 25 of 25

    A company has developed a new ML model. The company requires online model validation on 10% of the traffic before the company fully releases the model in production. The company uses an Amazon SageMaker endpoint behind an Application Load Balancer (ALB) to serve the model. Which solution will set up the required online validation with the LEAST operational overhead?

    • AUse production variants to add the new model to the existing SageMaker endpoint. Set the variant weight to 0.1 for the new model. Monitor the number of invocations by using Amazon CloudWatch.
    • BUse production variants to add the new model to the existing SageMaker endpoint. Set the variant weight to 1 for the new model. Monitor the number of invocations by using Amazon CloudWatch.
    • CCreate a new SageMaker endpoint. Use production variants to add the new model to the new endpoint. Monitor the number of invocations by using Amazon CloudWatch.
    • DConfigure the ALB to route 10% of the traffic to the new model at the existing SageMaker endpoint. Monitor the number of invocations by using AWS CloudTrail.

    Show answer

    Correct answer: A

    How others answered

    • A100%

    Explanation

    { 'ProductionVariants': [ { 'VariantName': 'existing-model', 'ModelName': 'existing-model', 'InitialVariantWeight': 0.9 }, { 'VariantName': 'new-model', 'ModelName': 'new-model', 'InitialVariantWeight': 0.1 } ] }

    Adapted from community discussion by GiorgioGss · 2024-11-28 · 2 community upvotes. Third-party contribution; reviewed by our quality filter, not independently verified.

    Supporting references

    • AWS documentation

    View community discussion3

    Candidate context

    Question 25 discussion

    These are quality-filtered excerpts from third-party community discussion. They can be useful reasoning signals, but they are not official answers and have not been independently verified by MLA-C01 or CertExamHQ.

    1. A

       aws_Tamilan

       2025-03-31

       Selected A1 upvote

       🔑 Keyword: Validate new model on 10% of traffic with minimal overhead ✅ Correct Answer: A. Use production variants to add the new model to the existing SageMaker endpoint. Set the variant weight to 0.1 for the new model. Monitor the number of invocations by using Amazon CloudWatch. Why? SageMaker production variants allow traffic splitting across multiple models on a single endpoint. Setting variant weight to 0.1 ensures only 10% of traffic is sent to the new model. CloudWatch can monitor invocations for validation. Why Others Are Wrong? ❌ B. Setting the variant weight to 1 will send all traffic to the new model, defeating the purpose. ❌ C. Creating a new endpoint increases operational overhead unnecessarily. ❌ D. ALB-based routing is more complex than using SageMaker variants for traffic splitting.

    2. G

       GiorgioGss

       2024-11-28

       Selected A2 upvotes

       { 'ProductionVariants': [ { 'VariantName': 'existing-model', 'ModelName': 'existing-model', 'InitialVariantWeight': 0.9 }, { 'VariantName': 'new-model', 'ModelName': 'new-model', 'InitialVariantWeight': 0.1 } ] }

    3. S

       Saransundar

       2024-12-04

       Selected A1 upvote

       https://docs.aws.amazon.com/sagemaker/latest/dg/model-ab-testing.html

Keyboard:J Kmove between questionsA–Dselect an answerRshow or hide the answer← →change page

MLA-C01 practice, question 1 of 25