1
AWS Certified Machine
Learning – Specialty
(MLS-C01 / MLA-C01)
Comprehensive Practice
Exam — 150 Multiple-
Choice Questions
, 2
Exam Overview: The AWS Certified Machine Learning – Specialty certification validates
expertise in building, training, tuning, and deploying machine learning models on AWS. The
exam covers four domains: Data Engineering (20%), Exploratory Data Analysis (24%), Modeling
(36%), and Machine Learning Implementation and Operations (20%). Candidates should have
two or more years of experience developing and running ML workloads in the AWS Cloud.
DOMAIN 1: DATA ENGINEERING (20%)
Questions 1–30
Question 1
A financial services company ingests 5 TB of daily transaction data from multiple sources into
Amazon S3. The data arrives in JSON, CSV, and log formats. A Machine Learning Specialist needs
to build a scalable data pipeline that automatically detects schema changes, converts data to
Parquet, partitions it by date and customer ID, and makes it available for Athena queries with
minimal latency. Which combination of AWS services should the Specialist use?
A. AWS Glue crawlers + AWS Glue ETL jobs + AWS Glue Data Catalog + Amazon S3
B. Amazon EMR with Spark + Apache Hive metastore + Amazon S3
C. AWS Lambda + Amazon Kinesis Data Firehose + Amazon S3 + Amazon Redshift Spectrum
D. AWS Database Migration Service + AWS Glue + Amazon RDS
- detailed answer 100% correct :- A
Rationale: AWS Glue crawlers automatically detect schema changes and populate the Glue Data
Catalog. Glue ETL jobs convert data to Parquet with partitioning. The Data Catalog enables
Athena querying. This serverless approach minimizes operational overhead and scales
automatically. Option B requires cluster management. Option C is for streaming, not batch.
Option D is for database migration.
Question 2
, 3
A Machine Learning Specialist needs to ingest real-time clickstream data from a mobile
application, perform light transformation (filtering and enrichment), and deliver the data to
Amazon S3 in Apache Parquet format for downstream model training. The solution must handle
millions of events per second with sub-second latency. Which service should the Specialist use?
A. Amazon Kinesis Data Streams with AWS Lambda consumers writing to S3
B. Amazon Kinesis Data Firehose with built-in Parquet conversion
C. AWS Glue streaming ETL with Parquet output
D. Amazon MSK with Kafka Connect S3 sink
- detailed answer 100% correct :- B
Rationale: Kinesis Data Firehose is the fully managed service designed for ingesting streaming
data and delivering it to S3. It supports native Parquet/ORC conversion and partitioning. It scales
automatically and requires minimal development effort. Option A requires managing Lambda
concurrency. Option C is batch-oriented. Option D requires Kafka cluster management.
Question 3
A data engineering team has 10 TB of historical data in Amazon S3 stored as uncompressed CSV
files. They need to transform this data using Apache Spark for feature engineering and store the
results in a columnar format optimized for analytical queries. The transformation runs weekly
and must complete within 2 hours. Which AWS service should the team use for the
transformation?
A. AWS Glue with Spark ETL jobs using G.1X or G.2X worker types
B. Amazon EMR with Spark on EC2 Spot Instances
C. AWS Batch with custom Spark containers
D. Amazon SageMaker Processing jobs with Spark
- detailed answer 100% correct :- B
Rationale: Amazon EMR with Spark provides the most cost-effective solution for large-scale
batch processing. Spot Instances reduce costs for fault-tolerant workloads. EMR can be
terminated after the job completes. AWS Glue workers may be insufficient for 10 TB within 2
hours. SageMaker Processing is for ML-specific preprocessing.
Question 4
, 4
A company stores IoT sensor data in Amazon S3. The data arrives in hourly partitions. The
Machine Learning Specialist needs to create a data catalog that enables analysts to query this
data using Amazon Athena without manually defining schemas. The solution should
automatically detect new partitions and schema evolution. Which AWS service should the
Specialist use?
A. AWS Glue Crawler scheduled to run hourly
B. AWS Glue Data Catalog with manual table definitions
C. Amazon Athena with MSCK REPAIR TABLE executed daily
D. AWS Lake Formation with automatic partition discovery
- detailed answer 100% correct :- A
Rationale: AWS Glue Crawlers automatically scan S3 data, infer schemas, populate the Data
Catalog, and detect new partitions. Scheduling the crawler hourly aligns with the data arrival
pattern. Option B requires manual effort. Option C doesn't detect schema evolution. Option D
requires Lake Formation setup without crawler capabilities.
Question 5
A Machine Learning Specialist is building a data pipeline that requires joining multiple large
datasets from Amazon S3 and Amazon RDS. The pipeline must run daily and complete within 30
minutes. Which service provides the most efficient way to perform these heterogeneous joins?
A. AWS Glue ETL with JDBC connection to RDS and S3 sources
B. Amazon EMR with Spark SQL reading from S3 and RDS
C. AWS Lambda with Pandas reading from S3 and RDS
D. Amazon Athena federated queries with RDS connector
- detailed answer 100% correct :- B
Rationale: Amazon EMR with Spark provides distributed processing optimized for large-scale
joins. Spark can efficiently read from both S3 (via S3A) and RDS (via JDBC) in a cluster. AWS Glue
may struggle with performance for very large joins. Lambda has memory/time limits. Athena
federated queries don't support complex joins well.
Question 6
A team stores training data in Amazon S3 with the following folder
structure: s3://bucket/year=2026/month=01/day=15/. They query this data using Amazon
AWS Certified Machine
Learning – Specialty
(MLS-C01 / MLA-C01)
Comprehensive Practice
Exam — 150 Multiple-
Choice Questions
, 2
Exam Overview: The AWS Certified Machine Learning – Specialty certification validates
expertise in building, training, tuning, and deploying machine learning models on AWS. The
exam covers four domains: Data Engineering (20%), Exploratory Data Analysis (24%), Modeling
(36%), and Machine Learning Implementation and Operations (20%). Candidates should have
two or more years of experience developing and running ML workloads in the AWS Cloud.
DOMAIN 1: DATA ENGINEERING (20%)
Questions 1–30
Question 1
A financial services company ingests 5 TB of daily transaction data from multiple sources into
Amazon S3. The data arrives in JSON, CSV, and log formats. A Machine Learning Specialist needs
to build a scalable data pipeline that automatically detects schema changes, converts data to
Parquet, partitions it by date and customer ID, and makes it available for Athena queries with
minimal latency. Which combination of AWS services should the Specialist use?
A. AWS Glue crawlers + AWS Glue ETL jobs + AWS Glue Data Catalog + Amazon S3
B. Amazon EMR with Spark + Apache Hive metastore + Amazon S3
C. AWS Lambda + Amazon Kinesis Data Firehose + Amazon S3 + Amazon Redshift Spectrum
D. AWS Database Migration Service + AWS Glue + Amazon RDS
- detailed answer 100% correct :- A
Rationale: AWS Glue crawlers automatically detect schema changes and populate the Glue Data
Catalog. Glue ETL jobs convert data to Parquet with partitioning. The Data Catalog enables
Athena querying. This serverless approach minimizes operational overhead and scales
automatically. Option B requires cluster management. Option C is for streaming, not batch.
Option D is for database migration.
Question 2
, 3
A Machine Learning Specialist needs to ingest real-time clickstream data from a mobile
application, perform light transformation (filtering and enrichment), and deliver the data to
Amazon S3 in Apache Parquet format for downstream model training. The solution must handle
millions of events per second with sub-second latency. Which service should the Specialist use?
A. Amazon Kinesis Data Streams with AWS Lambda consumers writing to S3
B. Amazon Kinesis Data Firehose with built-in Parquet conversion
C. AWS Glue streaming ETL with Parquet output
D. Amazon MSK with Kafka Connect S3 sink
- detailed answer 100% correct :- B
Rationale: Kinesis Data Firehose is the fully managed service designed for ingesting streaming
data and delivering it to S3. It supports native Parquet/ORC conversion and partitioning. It scales
automatically and requires minimal development effort. Option A requires managing Lambda
concurrency. Option C is batch-oriented. Option D requires Kafka cluster management.
Question 3
A data engineering team has 10 TB of historical data in Amazon S3 stored as uncompressed CSV
files. They need to transform this data using Apache Spark for feature engineering and store the
results in a columnar format optimized for analytical queries. The transformation runs weekly
and must complete within 2 hours. Which AWS service should the team use for the
transformation?
A. AWS Glue with Spark ETL jobs using G.1X or G.2X worker types
B. Amazon EMR with Spark on EC2 Spot Instances
C. AWS Batch with custom Spark containers
D. Amazon SageMaker Processing jobs with Spark
- detailed answer 100% correct :- B
Rationale: Amazon EMR with Spark provides the most cost-effective solution for large-scale
batch processing. Spot Instances reduce costs for fault-tolerant workloads. EMR can be
terminated after the job completes. AWS Glue workers may be insufficient for 10 TB within 2
hours. SageMaker Processing is for ML-specific preprocessing.
Question 4
, 4
A company stores IoT sensor data in Amazon S3. The data arrives in hourly partitions. The
Machine Learning Specialist needs to create a data catalog that enables analysts to query this
data using Amazon Athena without manually defining schemas. The solution should
automatically detect new partitions and schema evolution. Which AWS service should the
Specialist use?
A. AWS Glue Crawler scheduled to run hourly
B. AWS Glue Data Catalog with manual table definitions
C. Amazon Athena with MSCK REPAIR TABLE executed daily
D. AWS Lake Formation with automatic partition discovery
- detailed answer 100% correct :- A
Rationale: AWS Glue Crawlers automatically scan S3 data, infer schemas, populate the Data
Catalog, and detect new partitions. Scheduling the crawler hourly aligns with the data arrival
pattern. Option B requires manual effort. Option C doesn't detect schema evolution. Option D
requires Lake Formation setup without crawler capabilities.
Question 5
A Machine Learning Specialist is building a data pipeline that requires joining multiple large
datasets from Amazon S3 and Amazon RDS. The pipeline must run daily and complete within 30
minutes. Which service provides the most efficient way to perform these heterogeneous joins?
A. AWS Glue ETL with JDBC connection to RDS and S3 sources
B. Amazon EMR with Spark SQL reading from S3 and RDS
C. AWS Lambda with Pandas reading from S3 and RDS
D. Amazon Athena federated queries with RDS connector
- detailed answer 100% correct :- B
Rationale: Amazon EMR with Spark provides distributed processing optimized for large-scale
joins. Spark can efficiently read from both S3 (via S3A) and RDS (via JDBC) in a cluster. AWS Glue
may struggle with performance for very large joins. Lambda has memory/time limits. Athena
federated queries don't support complex joins well.
Question 6
A team stores training data in Amazon S3 with the following folder
structure: s3://bucket/year=2026/month=01/day=15/. They query this data using Amazon
