Skip to content

CLF-C02 Study Notes: AWS AI/ML and Analytics Services

1. Artificial Intelligence and Machine Learning Fundamentals

Artificial intelligence (AI)

AI is the field of computer science focused on solving tasks commonly associated with human intelligence, such as:

  • Learning
  • Problem solving
  • Pattern recognition
  • Understanding language
  • Recognizing images or speech

Machine learning (ML)

Machine learning uses algorithms and statistical models to identify patterns in historical data and make predictions or decisions without being explicitly programmed for every situation.

Common use cases include:

  • Fraud detection
  • Anomaly detection
  • Customer churn prediction
  • Product recommendations
  • Sentiment analysis
  • Image or video classification
  • Predicting future sales

2. AWS AI/ML Service Levels

AWS provides AI and ML capabilities at three general levels.

Level Description Typical AWS services
AI services Ready-to-use, fully managed AI capabilities accessed through APIs Amazon Rekognition, Amazon Lex, Amazon Polly, Amazon Translate
Machine learning services Tools to build, train, and deploy custom ML models Amazon SageMaker
ML frameworks and infrastructure Infrastructure and frameworks for users who need more control TensorFlow, PyTorch, Deep Learning AMIs, Deep Learning Containers, ML-optimized EC2 instances

Exam distinction

  • Use an AI service when you want to add a prebuilt AI capability to an application without developing and training a model.
  • Use Amazon SageMaker when you need to build, train, and deploy a custom machine learning model.
  • Use ML frameworks and infrastructure when you require greater control over the ML environment, frameworks, or compute resources.

3. AWS AI Services

AI services are generally pre-trained or automatically trained services that can be used through APIs. They allow developers to add AI capabilities without requiring extensive ML expertise.

Amazon Rekognition

Amazon Rekognition adds image and video analysis capabilities to applications.

It can help identify:

  • Objects
  • People
  • Faces
  • Text in images
  • Activities
  • Potentially unsafe content

Typical question wording

  • Analyze images or videos
  • Identify objects or people
  • Search, verify, and organize millions of images
  • Add computer vision capabilities to an application

Remember

Rekognition = image and video analysis

Amazon Translate

Amazon Translate provides machine translation for text.

Use it to:

  • Translate text between languages
  • Localize application content
  • Support multilingual applications

Remember

Translate = text translation

Amazon Polly

Amazon Polly converts text into lifelike speech.

Use it to:

  • Create audio from written text
  • Build voice-enabled applications
  • Add speech output to applications

Remember

Polly = text-to-speech

Amazon Lex

Amazon Lex is used to build conversational interfaces and chatbots.

It supports interactions using:

  • Voice
  • Text
  • Natural language

Remember

Lex = conversational chatbots

Amazon SageMaker

Amazon SageMaker is a managed machine learning service used to:

  • Build ML models
  • Train ML models
  • Deploy ML models
  • Scale ML workloads

SageMaker is appropriate when an organization wants to use its own data to create a custom model, such as a model that predicts future sales.

SageMaker versus AI services

Requirement Appropriate service
Analyze an image using a prebuilt API Amazon Rekognition
Translate text Amazon Translate
Convert text to speech Amazon Polly
Build a chatbot Amazon Lex
Build, train, and deploy a custom ML model Amazon SageMaker

Exam trap

Do not choose SageMaker merely because a question mentions machine learning. If the question asks for a specific prebuilt capability—such as image recognition, translation, or speech synthesis—choose the corresponding AI service.

ML Frameworks and Infrastructure

AWS supports open-source ML frameworks such as:

  • TensorFlow
  • PyTorch
  • Apache MXNet

AWS also provides:

  • Deep Learning AMIs: EC2 machine images with ML frameworks and tools preinstalled.
  • Deep Learning Containers: Container images optimized for ML workloads.
  • ML-optimized EC2 instances: Specialized compute for demanding ML workloads.

These options provide more control than fully managed AI services but require more technical knowledge and management.

4. Analytics Services

Data analytics converts raw data into useful information and insights. Organizations use analytics to:

  • Understand customer behavior
  • Improve operations
  • Analyze application activity
  • Build reports and dashboards
  • Identify trends
  • Support business decisions

Amazon Athena

Amazon Athena is a serverless interactive query service that uses standard SQL to analyze data stored in Amazon S3.

Key characteristics

  • Serverless: no infrastructure to provision or manage
  • Queries data directly in S3
  • Uses standard SQL
  • Charges based primarily on the amount of data scanned
  • Can query formats such as:
  • CSV
  • JSON
  • XML
  • Parquet

Common use cases

  • Querying log files in S3
  • Analyzing clickstream data
  • Generating reports
  • Analyzing AWS Cost and Usage Reports stored in S3

Remember

Athena = query data in S3 using SQL

Athena versus Amazon Redshift

Feature Amazon Athena Amazon Redshift
Infrastructure Serverless Data warehouse cluster or managed warehouse
Data location Queries data directly in S3 Primarily queries data loaded into Redshift
Best suited for Ad hoc queries and occasional analysis Large-scale, repeated analytical workloads
Pricing concept Pay based on data scanned Pay for provisioned or managed warehouse resources
Schema/data preparation Can query data in place with less preparation Typically requires structured warehouse data

Amazon Redshift

Amazon Redshift is a petabyte-scale data warehousing service designed for analytical workloads and online analytical processing (OLAP).

Key characteristics

  • Uses a column-based database engine
  • Designed for large-scale analytics
  • Uses a cluster-based architecture
  • Can load data from sources such as:
  • Amazon S3
  • Amazon DynamoDB
  • Amazon Kinesis
  • Other databases using migration or integration services

Redshift Spectrum

Redshift Spectrum allows a Redshift data warehouse to query data directly in S3 without first loading all of it into Redshift.

Redshift Spectrum versus Athena

  • Athena is serverless and does not require a Redshift cluster.
  • Redshift Spectrum requires Redshift, but allows the Redshift environment to query external data in S3.

Exam trap

If the question emphasizes a petabyte-scale data warehouse or OLAP, select Amazon Redshift—not Athena.

Amazon Macie

Amazon Macie is a security service that uses machine learning to discover, classify, and help protect sensitive data stored in Amazon S3.

It can help identify personally identifiable information (PII), such as:

  • Names
  • Addresses
  • Email addresses
  • Government identification numbers
  • Passport or driver’s license numbers
  • Dates of birth
  • Bank account information
  • Credit card information

Macie provides dashboards, reports, and alerts about sensitive data.

It can also analyze AWS CloudTrail logs.

Remember

Macie = sensitive data and PII discovery in S3

Exam memory aid

“Macie” contains an i, as does PII.

Macie versus Athena

Requirement Appropriate service
Run SQL queries against data in S3 Amazon Athena
Discover sensitive data or PII in S3 Amazon Macie

Exam trap

Although Macie uses machine learning, it is primarily a security and data discovery service, not a general-purpose ML model-building service.

Amazon Kinesis

Amazon Kinesis is used to collect, process, and analyze streaming data in real time and at scale.

Examples of streaming data include:

  • Application logs
  • Website clickstreams
  • Video
  • Audio
  • IoT data
  • Mobile application activity

Common question wording

  • Real-time data ingestion
  • Streaming data
  • Processing millions of events as they occur
  • Analyzing clickstream or IoT data immediately

Remember

Kinesis = real-time streaming data

Kinesis versus batch analytics

  • Use Kinesis for continuously arriving, real-time data.
  • Use services such as Athena for querying data that is already stored in S3.

Kinesis services

The Kinesis family includes services such as:

  • Kinesis Data Firehose: Delivers streaming data to destinations such as S3 and other AWS services.
  • Kinesis Data Analytics: Analyzes streaming data in real time.

For the CLF-C02 exam, focus on the broad distinction that Kinesis handles streaming data.

AWS Glue

AWS Glue is a serverless data integration service.

It helps organizations:

  • Discover data
  • Prepare and clean data
  • Move and integrate data
  • Create extract, transform, and load (ETL) pipelines
  • Prepare data for analytics and machine learning
  • Catalog data for later querying

Glue can help load and prepare data for data lakes. Cataloged data can then be queried using services such as:

  • Amazon Athena
  • Amazon Redshift Spectrum
  • Amazon EMR

Remember

Glue = serverless ETL and data integration

Exam trap

Glue is not primarily a visualization or dashboard service. It prepares and integrates data so other analytics services can use it.

Amazon QuickSight

Amazon QuickSight is a fully managed business intelligence (BI) service.

It can be used to:

  • Create interactive dashboards
  • Visualize data
  • Share business insights
  • Provide reports to users across an organization
  • Include machine learning-powered insights

Remember

QuickSight = business intelligence dashboards and visualization

QuickSight versus AWS Glue

Requirement Appropriate service
Prepare and transform data AWS Glue
Create dashboards and visualizations Amazon QuickSight

Amazon EMR

Amazon EMR is a managed service for processing and analyzing very large datasets.

It uses frameworks such as Apache Hadoop and Apache Spark and can support:

  • Big data processing
  • Data transformation
  • Analytics
  • Machine learning
  • Log analysis
  • Scientific simulations
  • Bioinformatics
  • Deep learning workloads

EMR uses AWS infrastructure such as Amazon EC2 and can work with data in Amazon S3.

Remember

EMR = managed big data processing using frameworks such as Hadoop and Spark

EMR versus QuickSight

  • EMR processes and transforms large datasets.
  • QuickSight creates business intelligence dashboards and visualizations.

Exam trap

If the question mentions Apache Spark, Apache Hadoop, or large-scale data transformation, Amazon EMR is a likely answer—not QuickSight or OpenSearch.

5. Important Service Comparisons

AI service versus SageMaker

Scenario Service
Add image and video analysis to an application Amazon Rekognition
Convert text into speech Amazon Polly
Translate text Amazon Translate
Build a chatbot Amazon Lex
Build and deploy a custom ML model Amazon SageMaker

Analytics service selection

Scenario Service
Query data in S3 using SQL Amazon Athena
Data warehouse for large-scale OLAP Amazon Redshift
Query S3 data from Redshift without loading it Redshift Spectrum
Detect PII and sensitive data in S3 Amazon Macie
Process real-time streaming data Amazon Kinesis
Perform serverless ETL and data integration AWS Glue
Create BI dashboards Amazon QuickSight
Process large datasets with Hadoop or Spark Amazon EMR

6. Exam Tips and Traps

  1. “Image or video analysis” means Rekognition.
    Do not select SageMaker unless the question asks you to build a custom model.

  2. “Text-to-speech” means Polly.

  3. “Translate text” means Translate.

  4. “Conversational chatbot” means Lex.

  5. “Build, train, and deploy a custom ML model” means SageMaker.

  6. “Query data in S3 using SQL” means Athena.

  7. “PII or sensitive data in S3” means Macie.

  8. “Real-time streaming data” means Kinesis.

  9. “ETL, data catalog, or data integration” means AWS Glue.

  10. “Dashboards and business intelligence” means QuickSight.

  11. “Apache Spark or Hadoop for big data processing” means EMR.

  12. Athena and Redshift are not the same:

    • Athena is serverless and queries data directly in S3.
    • Redshift is a data warehouse designed for large-scale analytical workloads.

  13. Redshift Spectrum still requires Redshift.
    It is not a completely serverless alternative to Athena.

  14. Macie is primarily a security service.
    Its use of machine learning does not make it a replacement for SageMaker.

7. Practice Questions

Question 1

A company wants to add the ability to identify objects and people in uploaded images without developing and training its own machine learning model. Which AWS service should it use?

A. Amazon SageMaker
B. Amazon Rekognition
C. Amazon Polly
D. Amazon Athena

Answer: B. Amazon Rekognition

Explanation: Rekognition provides prebuilt image and video analysis capabilities through APIs.

Question 2

Which AWS service converts text into lifelike speech?

A. Amazon Lex
B. Amazon Translate
C. Amazon Polly
D. Amazon Rekognition

Answer: C. Amazon Polly

Explanation: Amazon Polly is a text-to-speech service.

Question 3

A company needs to analyze log files stored in Amazon S3 using standard SQL. The company does not want to provision or manage servers. Which service should it use?

A. Amazon Athena
B. Amazon Redshift
C. Amazon EMR
D. Amazon QuickSight

Answer: A. Amazon Athena

Explanation: Athena is a serverless interactive query service that analyzes data directly in S3 using SQL.

Question 4

Which AWS service helps discover and classify personally identifiable information stored in Amazon S3?

A. Amazon Athena
B. Amazon Macie
C. AWS Glue
D. Amazon Kinesis

Answer: B. Amazon Macie

Explanation: Macie uses machine learning to discover and classify sensitive data, including PII, in S3.

Question 5

A mobile application generates millions of clickstream events that must be ingested and processed as they occur. Which AWS service is most appropriate?

A. Amazon Kinesis
B. Amazon Redshift
C. Amazon QuickSight
D. Amazon S3 Glacier

Answer: A. Amazon Kinesis

Explanation: Kinesis is designed for real-time ingestion and processing of streaming data.

Question 6

A company needs a serverless service to discover, prepare, and transform data from multiple sources before loading it into a data lake. Which service should it use?

A. Amazon QuickSight
B. AWS Glue
C. Amazon Lex
D. Amazon Rekognition

Answer: B. AWS Glue

Explanation: AWS Glue is a serverless data integration and ETL service.

Question 7

Which AWS service is designed to create and publish interactive business intelligence dashboards?

A. Amazon EMR
B. Amazon Athena
C. Amazon QuickSight
D. Amazon SageMaker

Answer: C. Amazon QuickSight

Explanation: QuickSight is a managed business intelligence and visualization service.

Question 8

A data engineering team needs to use Apache Spark to perform large-scale data transformations and analytics. Which AWS service should it use?

A. Amazon EMR
B. Amazon Macie
C. Amazon Polly
D. Amazon QuickSight

Answer: A. Amazon EMR

Explanation: Amazon EMR provides managed big data processing using frameworks such as Apache Spark and Hadoop.

Question 9

A company wants to develop, train, and deploy a custom machine learning model using its own historical sales data. Which service is most appropriate?

A. Amazon Translate
B. Amazon SageMaker
C. Amazon Rekognition
D. Amazon Kinesis

Answer: B. Amazon SageMaker

Explanation: SageMaker is used to build, train, and deploy custom machine learning models.

Question 10

Which statement correctly compares Amazon Athena and Amazon Redshift?

A. Athena is used for chatbots, while Redshift is used for translation.
B. Athena queries data in S3 without requiring a data warehouse, while Redshift is a data warehouse for analytical workloads.
C. Athena is used for real-time streaming, while Redshift is used for PII discovery.
D. Athena creates dashboards, while Redshift performs text-to-speech conversion.

Answer: B. Athena queries data in S3 without requiring a data warehouse, while Redshift is a data warehouse for analytical workloads.