PracticeCF

MLA-C01 Question Bank · Latest 50 Questions Free Preview (Page 1/5)

Questions 1–10 of the latest 50 MLA-C01 questions (5 pages total), with answers and explanations, no login required. Practice all 233 questions online.

Question 1 · No. 184 · Single choice

A company is developing a new ML model that uses the XGBoost algorithm. The company will train the model on data that is stored in an Amazon S3 bucket. The data is in a nested JSON format. An ML engineer needs to convert the JSON files into a tabular format. Which solution will meet this requirement with the LEAST operational overhead?

Answer: A

Explanation: AWS Glue provides a built-in transform called 'Relationalize' that is specifically designed to flatten nested JSON structures into a tabular format suitable for machine learning workflows. Using an AWS Glue PySpark job leverages a managed service, minimizing operational overhead compared to writing and managing custom code (as in options B and C) or relying on Athena's limited transformation capabilities (option D). Athena can query nested JSON but does not natively support a 'flatten' function that fully relationalizes deeply nested data like Glue’s Relationalize transform does. Therefore, option A offers the least operational overhead while effectively meeting the requirement.
Question 2 · No. 185 · Single choice

A company is using an Amazon S3 bucket to collect data that will be used for ML workflows. The company needs to use AWS Glue DataBrew to clean and normalize the data. Which solution will meet these requirements?

Answer: B

Explanation: AWS Glue DataBrew supports creating datasets directly from Amazon S3 using an S3 path. To clean and normalize data, DataBrew uses 'recipe jobs,' which apply a series of data transformation steps defined in a recipe. Profile jobs are used for analyzing and understanding data (e.g., generating statistics), not for performing transformations. Additionally, JDBC drivers are not used to connect to S3 buckets, as S3 is an object storage service, not a relational database. Therefore, option B is the correct solution.
Question 3 · No. 186 · Single choice

A company stores training data as a .csv file in an Amazon S3 bucket. The company must encrypt the data and must control which applications have access to the encryption key. Which solution will meet these requirements?

Answer: D

Explanation: The requirement specifies that the data must be encrypted and the company must control which applications have access to the encryption key. AWS KMS (Key Management Service) is designed exactly for this purpose—it allows centralized control over encryption keys and fine-grained access management via IAM policies. Option D correctly proposes creating a new AWS KMS key and using the AWS Encryption CLI with that key to encrypt the file. This ensures both encryption of data and controlled access to the key. Other options are incorrect: SSH keys (A) are not used for data encryption in S3; API keys from API Gateway (B) are for API authentication, not encryption; and while IAM roles (C) can grant permissions to use KMS keys, the option does not mention creating a KMS key itself, which is essential for encryption.
Question 4 · No. 187 · Single choice

A company wants to build a real-time analytics application that uses streaming data from social media. An ML engineer must implement a solution that ingests and transforms 5 GB of data each minute. The solution also must load the data into a data store that supports fast queries for the real-time analytics. Which solution will meet these requirements?

Answer: D

Explanation: Option D is the best choice because it uses Amazon Kinesis Data Streams, which is designed for real-time ingestion of high-volume streaming data (5 GB per minute). Amazon Managed Service for Apache Flink (formerly Kinesis Data Analytics) is purpose-built for real-time stream processing and transformation. Finally, Amazon DynamoDB is a fully managed NoSQL database that supports fast, low-latency queries—ideal for real-time analytics applications. The other options use services not optimized for high-throughput real-time streaming: EventBridge, SQS, and SNS are not designed for continuous high-volume data ingestion; ElastiCache is an in-memory cache, not a persistent queryable data store; S3 is object storage not optimized for fast interactive queries; and RDS is a relational database that may not scale as effectively for high-velocity, schema-flexible streaming workloads.
Question 5 · No. 188 · Single choice

A company has an existing Amazon SageMaker model (v1) on a production endpoint. The company develops a new model version (v2) and needs to test v2 in production before substituting v2 for v1. The company needs to implement a solution to minimize the risk of v2 generating incorrect output in production. The solution must prevent any disruption of production traffic during the change to v2. Which solution will meet these requirements?

Answer: D

Explanation: Option D uses SageMaker's shadow variant feature, which allows v2 to receive 100% of the production traffic copies without affecting the actual responses sent to clients (which still come from v1). This enables thorough testing of v2 under real production conditions while ensuring zero disruption to users. All outputs from v2 can be captured and analyzed for correctness before promoting it to production. Options A and B use traffic splitting, which risks sending potentially incorrect predictions from v2 to real users (even if only 1% or 10%). Option C deploys v2 to a separate endpoint and manually reroutes traffic, which doesn't guarantee identical input conditions and introduces operational complexity and potential disruption during the switch. Therefore, D is the safest and most effective approach.
Question 6 · No. 189 · Single choice

A company is building an ML model by using Amazon SageMaker, AWS owned libraries, and open source libraries. The company must ensure that SageMaker does not collect metadata about usage and errors during training. Which solution will meet these requirements?

Answer: D

Question 7 · No. 190 · Single choice

An ML engineer is training an ML model to identify people’s health risk based on 20 features and 1 target. The target class has two values: • Likely to have health risk (positive class) • Unlikely to have health risk (negative class) The age range of people in the dataset is 30 years old to 60 years old. Age is one of the features. The ML engineer analyzes the features. For the positive class, the difference in proportions of labels (DPL) value is (+0.9) for the age range of 40 to 45 compared with all other age ranges. What should the ML engineer do to correct this data imbalance?

Answer: B

Explanation: The DPL (Difference in Proportions of Labels) value of +0.9 for the age range 40–45 in the positive class indicates a significant overrepresentation of positive cases in that subgroup compared to others. This suggests bias or imbalance where the model may become overly confident or skewed toward predicting positive outcomes for people aged 40–45. To correct this imbalance, the engineer should reduce the dominance of this subgroup by undersampling the positive class specifically for the 40–45 age range, which corresponds to option B.
Question 8 · No. 191 · Multiple choice

A company wants to launch a new internal generative AI interface to answer user questions. The interface will be based on a popular open source large language model (LLM). Which combination of steps will deploy the interface with the LEAST operational overhead? (Choose two.)

Answer: A, E

Question 9 · No. 192 · Single choice

A company runs an ML model on Amazon SageMaker. The company uses an automatic process that makes API calls to create training jobs for the model. The company has new compliance rules that prohibit the collection of aggregated metadata from training jobs. Which solution will prevent SageMaker from collecting metadata from the training jobs?

Answer: A

Explanation: The question asks for a solution to prevent Amazon SageMaker from collecting aggregated metadata from training jobs. Option A directly addresses this by opting out of metadata tracking, which is a feature SageMaker provides to disable the collection of usage metrics and metadata for compliance or privacy reasons. The other options—using a private subnet (B), encrypting data with KMS (C), or using Nitro instances (D)—are related to network isolation, data security, and hardware features respectively, but none of them stop SageMaker from collecting metadata about the training job itself.
Question 10 · No. 193 · Single choice

A company needs to use Amazon SageMaker to train a model on more than 300 GB of data. The training data is composed of files that are 200 MB in size. The data is stored in Amazon S3 Standard storage and feeds a dashboard tool. Which SageMaker training ingestion mechanism is the MOST cost-effective solution for this scenario?

Answer: D

12345 Next