What is Amazon Rekognition and what can it do?
Fully managed computer vision service.
Performs image and video analysis without needing ML expertise.
Supports:
Object & scene detection
Face detection & analysis (attributes, emotions)
Face comparison & search
Text detection (OCR)
Content moderation
Celebrity recognition
Integrates with S3, Lambda, Kinesis Video Streams, and other AWS services.
What does Amazon Rekognition’s Content Moderation feature do?
Automatically detects inappropriate, unsafe, or offensive content in images and videos.
Identifies categories such as explicit, suggestive, violence, weapons, and more.
Returns confidence scores for each moderation label.
Helps enforce platform safety policies for user-generated content.
Commonly used in social media, e-commerce, and content platforms.
What is Amazon Transcribe and what is it used for?
Fully managed automatic speech-to-text service.
Converts audio and video into accurate, searchable text.
Supports features like:
Speaker identification
Custom vocabulary
Automatic punctuation
Timestamp generation
Language identification
Common use cases: call center analytics, subtitles, meeting transcripts, and voice-driven apps.
What is Amazon Polly and what is it used for?
Fully managed text-to-speech (TTS) service.
Converts written text into natural-sounding audio.
Supports dozens of voices and languages.
Offers Neural TTS (NTTS) for more human-like speech.
Can generate streaming audio for real-time applications.
Common use cases: voice assistants, narration, accessibility, IVR systems, and audio content generation.
How do Amazon Polly Lexicons and SSML enhance text-to-speech output?
Lexicons
Let you customize pronunciation using phonetic rules.
Useful for brand names, acronyms, or domain-specific terms.
SSML (Speech Synthesis Markup Language)
Adds control over prosody, emphasis, pauses, volume, speed, and pitch.
Enables expressive, natural-sounding speech output.
Together:
Enable fine-grained control over how Polly speaks text.
What is Amazon Translate and what is it used for?
Fully managed neural machine translation service.
Translates text between dozens of languages with high accuracy.
Supports real-time and batch translation.
Provides custom terminology to preserve domain-specific vocabulary.
Common use cases: multilingual apps, content localization, chat translation, and document processing.
What are Amazon Lex and Amazon Connect used for?
Amazon Lex
Service for building conversational interfaces using voice or text.
Uses the same NLP technology as Alexa.
Handles intents, slots, and dialog management.
Amazon Connect
Cloud-based contact center service.
Integrates with Lex to provide automated voice/chat interactions.
Supports call routing, analytics, and customer experience workflows.
Together:
Enable intelligent, automated customer support with natural language interactions.
What is Amazon Comprehend and what does it do?
Fully managed Natural Language Processing (NLP) service.
Uses machine learning to extract insights from text.
Capabilities include:
Entity recognition (people, places, organizations)
Sentiment analysis
Key phrase extraction
Language detection
PII detection & redaction
Topic modeling
Ideal for analyzing documents, support tickets, reviews, and large text datasets.
What is Amazon Comprehend Medical and what does it extract from medical text?
Specialized NLP service for healthcare and life sciences.
Automatically identifies medical information such as:
Medical conditions
Medications & dosages
Anatomy terms
Test names & results
Treatments & procedures
Helps process clinical notes, EHRs, insurance documents, and medical transcripts.
Designed to improve accuracy over general-purpose NLP models.
What is Amazon SageMaker and what is it used for?
Fully managed service for building, training, and deploying machine learning models at scale.
Provides built-in algorithms, notebooks, autopilot ML, and feature engineering tools.
Handles infrastructure, distributed training, MLOps, and auto-scaling.
Supports real-time, batch, and asynchronous inference.
Integrates with S3, ECR, Lambda, Glue, and other AWS data/ML services.
What is Amazon Kendra and what is it used for?
Fully managed enterprise search service powered by machine learning.
Delivers highly accurate, natural-language search across organizational content.
Connects to many data sources: S3, SharePoint, Salesforce, Confluence, RDS, and more.
Supports semantic search, FAQ matching, and intelligent ranking.
Ideal for knowledge bases, internal portals, and customer support search.
What is Amazon Personalize and what does it provide?
Fully managed ML service for real-time personalization and recommendations.
Uses the same technology Amazon.com uses for product recommendations.
Supports:
User personalization
Item-to-item recommendations
Personalized search
Context-aware recommendations (time, device, etc.)
Requires only interaction data (and optionally item/user metadata).
No ML expertise required—service handles training, tuning, and deployment.
What is Amazon Textract and what can it extract?
Fully managed OCR and document extraction service.
Extracts:
Printed text
Handwriting
Tables
Forms and key-value pairs
Works with PDFs, scanned documents, images, and forms.
Ideal for automating data entry, document processing, and compliance workflows.
What are the key AWS services in the Machine Learning ecosystem?
AI Services (no ML experience needed):
Rekognition (vision)
Transcribe (speech-to-text)
Polly (text-to-speech)
Translate (language translation)
Comprehend (NLP)
Textract (document extraction)
Personalize (recommendations)
Forecast (time-series forecasting)
Kendra (enterprise search)
ML Services:
SageMaker for building, training, and deploying ML models
Data + Analytics Foundation:
S3, Glue, Athena, Redshift, EMR, Kinesis/MSK for ingesting and preparing data
Big idea:
AWS ML spans AI APIs, custom ML with SageMaker, and data services that power ML pipelines.
