MLA-C02PRACTICE QUESTIONS

MLA-C02 practice questions with answers

4 example questions from the AWS Certified Machine Learning Engineer - Associate question bank, grouped by exam domain. Pick an answer and check it to see the explanation.

Data Preparation for ML and AI

Question 1Select 1

null Collect and store data

A retailer stores transaction records in Amazon RDS for MySQL. An ML team needs an initial copy of the records in Amazon S3 and an ongoing feed of committed inserts, updates, and deletes for feature preparation. After the initial load, the feed must preserve successive changes to the same record, even when those changes occur between extraction runs. The application permanently deletes some records and cannot be modified to produce change events. Which solution meets these requirements with the least custom development?

A. Schedule an AWS Glue JDBC job that uses the transaction ID as a job bookmark and appends newly extracted rows to Amazon S3.

B. Enable the required row-based binary logging and configure sufficient log retention on the database. Use an AWS DMS full-load-and-CDC task with the database as the source and Amazon S3 as the target.

C. Create an RDS read replica and periodically query rows whose last_modified timestamp exceeds the previous extraction watermark. Append the query results to Amazon S3.

D. Take recurring RDS snapshots, export each snapshot to Amazon S3, and use an AWS Glue job to compare consecutive exports and generate change records.

ML Model and Foundation Model (FM) Development

Question 1Select 1

null Analyze and evaluate the performance of ML and AI systems

A machine learning engineer runs a TensorFlow training job on an Amazon EC2 GPU instance using an AWS Deep Learning Container. Each epoch contains 200 batches, and the optimizer performs one update per batch. The engineer intends to halve the learning rate every 5 epochs during a 20-epoch run. The training script uses tf.keras.optimizers.schedules.ExponentialDecay(initial_learning_rate=0.001, decay_steps=5, decay_rate=0.5, staircase=False). Training loss improves during the first few batches but then stops improving. Logs show that the learning rate becomes nearly zero during the first epoch, while gradients remain finite and nonzero. Which change best addresses the convergence issue while implementing the intended decay interval?

A. Configure the optimizer with global_clipnorm=1.0 and retain the existing learning rate schedule.

B. Set decay_steps=1000 and retain the initial learning rate and decay rate.

C. Keep decay_steps=5 and set staircase=True so that the learning rate changes only at discrete intervals.

D. Multiply initial_learning_rate by 200 and keep decay_steps=5 to compensate for the number of batches per epoch.

Deployment and Orchestration of ML and AI Workflows

Question 1Select 1

null Manage deployment infrastructure for ML and AI model types

A media company must run a trained video-classification model against 1,200 independent input shards in Amazon S3 each weekend. Each shard requires one NVIDIA GPU and approximately 75 minutes of processing. Results are needed the next day, not interactively. The existing container runs a command-line inference program that reads a shard and writes predictions to Amazon S3; it does not implement an HTTP inference server. The company cannot modify the container and wants AWS to provision and scale the compute capacity, with no running compute instances between workloads. Which solution meets these requirements?

A. Create an AWS Batch unmanaged EC2 compute environment with GPU-capable instance types. Associate it with a job queue and rely on the job queue to launch instances when array-job children are waiting and terminate them when processing finishes.

B. Register the existing container as an Amazon SageMaker AI model and launch a GPU-backed Batch Transform job against the S3 input shards, using the container unchanged.

C. Create an AWS Batch managed Fargate compute environment. Define a job that requests one GPU and submit an array job whose children each process one shard.

D. Create an AWS Batch managed EC2 compute environment with GPU-capable instance types and minimum vCPUs set to zero. Associate it with a job queue, define a job that requests one GPU, and submit an array job whose children each process one shard.

Operating, Monitoring, and Securing ML and AI Solutions

Question 1Select 1

null Optimize and manage ML and AI infrastructure costs and performance

A company is selecting an Amazon EC2 instance configuration for a synchronous image-classification API. Traffic alternates between sparse requests and sustained bursts. The API must return a complete response within 100 milliseconds for at least 99% of requests during both traffic patterns. An ML engineer tests four configurations using representative requests. Each configuration meets the required throughput and model accuracy. The client-observed latency includes network transport, preprocessing, batch waiting, model execution, and response delivery. The table shows the higher p99 latency observed across the sparse-traffic and burst-traffic tests. Prices are illustrative. | Configuration | Hourly cost per instance | p99 model execution latency | p99 client-observed latency | |---|---:|---:|---:| | Compute-optimized CPU, no batching | $0.60 | 85 ms | 125 ms | | Inference accelerator, batches of up to 32 requests | $0.85 | 18 ms | 145 ms | | Small GPU instance, no batching | $1.20 | 42 ms | 78 ms | | Large GPU instance, no batching | $2.40 | 24 ms | 58 ms | The company will use the same fixed instance count for every configuration. It must select one of the tested configurations without further serving changes. Which choice meets the response-time requirement at the lowest instance cost?

A. Select the compute-optimized CPU configuration because its model execution p99 is below 100 milliseconds and it has the lowest hourly price.

B. Select the small GPU configuration because its client-observed p99 latency meets the deadline and its hourly cost is the lowest among compliant configurations.

C. Select the large GPU configuration because its lower client-observed p99 latency makes it the lowest-cost configuration that satisfies the response-time requirement.

D. Select the inference-accelerator configuration because it has the lowest model execution latency and costs less than either GPU configuration.

PRACTICE EXAMMLA-C02 exam simulator

861 questions covering every domain and task, full-length exams, explanations for every answer and a pass probability indicator.

Open the simulator →