CLF-C02
Deep Dive
AWS offers pre-trained AI/ML services (no ML knowledge required) as well as platforms for custom ML. The CLF-C02 tests which service fits each use case: images, text, voice, search, or data analysis.
Contents
Pre-built AI services
High-level services that add AI capabilities directly to your applications. You don't need to know ML or train models. Just call an API and get the result.
Amazon SageMaker — Custom ML
For when the pre-built services don't fit your use case and you need to train your own ML models. SageMaker manages the entire ML infrastructure.
Exam decision rule
Use pre-built services (Rekognition, Lex, Comprehend, etc.) when the use case is standard. Use SageMaker when you need to train a custom model on your own data and the problem isn't solved by the pre-built services.
Amazon SageMaker AI
Custom MLA complete platform for building, training, and deploying Machine Learning models. Includes managed Jupyter notebooks, distributed training, AutoML, and inference endpoints.
Amazon Rekognition
Computer visionImage and video analysis with ML. Detects objects, scenes, activities, faces (identification and comparison), inappropriate content, and text in images.
Amazon Comprehend
Natural language processing (NLP)Extracts insights and relationships from unstructured text. Sentiment analysis (positive/negative/neutral), entity extraction (people, places, dates), language detection, topic classification.
Amazon Translate
TranslationHigh-quality neural machine translation across more than 75 language pairs. Ideal for content localization, multilingual communication, and analyzing text in multiple languages.
Amazon Polly
Text to speechConverts text into lifelike speech (Text-to-Speech / TTS). Offers natural-sounding voices across dozens of languages and dialects, including more expressive neural voices.
Amazon Transcribe
Speech to textAutomatic speech recognition (ASR). Converts audio to text with high accuracy. Supports multiple languages, speaker identification, and custom vocabulary.
Amazon Lex
Chatbots / Conversational AIA service for building conversational interfaces (chatbots and voice assistants) using the same technology as Amazon Alexa. Understands intents and manages multi-turn conversations.
Amazon Kendra
Intelligent enterprise searchAn enterprise search engine powered by ML. Searches across documents, FAQs, knowledge bases, and multiple data repositories, returning direct answers rather than just a list of links.
Amazon Textract
Document extractionAutomatically extracts text, tables, and form data from scanned documents. Goes beyond simple OCR — it understands document structure (forms, tables, fields).
Amazon Q
Generative AI assistantA generative AI assistant for the enterprise. Amazon Q Business: an assistant connected to internal enterprise data (now part of Amazon Quick). Amazon Q Developer: an assistant for developers built into the IDE; since May 2026 it no longer accepts new sign-ups and its successor is Kiro.
Analytics services let you process, transform, and visualize large volumes of data.
| Service | Main function | Typical use case | Model |
|---|---|---|---|
Amazon Athena | Serverless SQL queries over data in S3 | "Query S3 access logs with SQL without moving the data" | Pay per query (per TB scanned) |
Amazon Kinesis | Real-time data processing (streaming) | Real-time clickstream analysis, IoT metrics, application logs | Kinesis Streams + Firehose |
AWS Glue | Serverless ETL (Extract, Transform, Load) | Prepares data from multiple sources for analysis: cleans, transforms, and loads into Redshift or S3 | Serverless, pay per use |
Amazon EMR | Big data with Hadoop/Spark on managed clusters | Processing petabytes of logs with Apache Spark for ML or large-scale analysis | On-demand EC2 clusters |
Amazon OpenSearch | Search and analytics for logs and data | Application log analysis, full-text document search, operational dashboards | Managed instances |
Amazon Redshift | Petabyte-scale data warehouse for analytics (OLAP) | Historical sales reports, Business Intelligence, combined with Quick Sight | Managed clusters or Serverless |
Amazon Quick Sight (formerly QuickSight)
Amazon Quick Sight (formerly QuickSight, now part of Amazon Quick) is AWS's serverless Business Intelligence (BI) service. It builds interactive dashboards and visualizations connected to Redshift, S3 (via Athena), RDS, and other sources. There's no icon for it in the icon library, but it appears often on the exam: QuickSight = dashboards and data visualization.
The exam presents scenarios where you must identify the correct AI service. Memorize these associations:
Image analysis / object detection / facial recognition
Amazon Rekognition
Chatbot / virtual assistant / conversational AI
Amazon Lex
Intelligent search across enterprise documents
Amazon Kendra
Extract text/tables/data from scanned documents or PDFs
Amazon Textract
Sentiment analysis / entities in text
Amazon Comprehend
Transcribe audio / recorded calls to text
Amazon Transcribe
Convert text to speech (TTS) / narrate content
Amazon Polly
Translate text between languages
Amazon Translate
Train a custom model / your own ML
Amazon SageMaker
Generative AI assistant for developers / AWS
Amazon Q Developer (successor: Kiro)
Serverless SQL queries over data in S3
Amazon Athena
Streaming data / real time / IoT
Amazon Kinesis
Machine Learning / AI
ML models predict, classify, or generate new information based on patterns learned from historical data.
Services: SageMaker, Rekognition, Lex, Comprehend, Polly, Transcribe, Textract
Data analytics
Analytics services query, transform, and visualize existing data to surface descriptive insights.
Services: Athena, Kinesis, Glue, Redshift, EMR, OpenSearch, Quick Sight
Common exam traps
• Redshift: it's analytics (OLAP data warehouse), NOT ML. Used for reporting and BI.
• Athena: SQL over S3. Analytics, not ML. Serverless and fast for data exploration.
• Comprehend vs Textract: Comprehend analyzes the MEANING of text (sentiment, entities). Textract EXTRACTS text from scanned documents (advanced OCR).
• Transcribe vs Polly: Transcribe = audio to text. Polly = text to audio. They're opposites.
Did you get this topic?
Test what you just learned
An e-commerce company wants to add a virtual assistant to its website to answer customer questions about orders, returns, and products without building ML models from scratch. Which AWS service should they use?
Put what you learned to the test with original CLF-C02 exam questions.
Start CLF-C02 mock examUnlimited mock exams with the Pro plan
Log in to track your progress.