AWS AI Services (Rekognition, Polly, Lex & Comprehend)
7 questions found
What is Amazon Rekognition and what kinds of tasks can it perform on images and videos?
Beginner
Amazon Rekognition is a fully managed computer vision service that can analyze images and videos to detect objects, scenes, faces, and text, and it can also perform tasks such as facial comparison, content moderation to flag inappropriate images, and celebrity recognition, all through a simple API call without requiring you to build or train your own computer vision models.
aws rekognition detect-labels --image '{"S3Object":{"Bucket":"my-bucket","Name":"photo.jpg"}}' --max-labels 10
Real-world example
A social media platform uses Rekognition to automatically scan uploaded images for inappropriate content, flagging anything that violates community guidelines before it becomes publicly visible to other users.
Common follow-ups: How accurate is Rekognition's facial comparison feature?;Can Rekognition be trained to recognize custom objects specific to a business?
S3 & Storage;IAM
What is Amazon Polly and how does it convert text into natural sounding speech?
Beginner
Amazon Polly is a text to speech service that uses advanced deep learning technology to turn written text into lifelike spoken audio in dozens of languages and voices, letting developers add voice capabilities such as reading articles aloud or providing voice responses in an application without needing any expertise in speech synthesis.
aws polly synthesize-speech --text 'Welcome to our application' --output-format mp3 --voice-id Joanna output.mp3
Real-world example
A news application uses Polly to automatically generate an audio version of every published article, allowing users to listen to the news during their commute instead of only reading it on screen.
Common follow-ups: What is the difference between standard voices and neural voices in Polly?;Can Polly speak text using Speech Synthesis Markup Language for finer control over pronunciation?
Amazon SageMaker & Machine Learning on AWS;S3 & Storage
What is Amazon Lex and how does it enable building conversational chatbots?
Intermediate
Amazon Lex is a service for building conversational interfaces, such as chatbots and voice assistants, using the same natural language understanding and automatic speech recognition technology that powers Amazon Alexa, letting you define intents representing what a user wants to accomplish, along with sample phrases and slots for capturing specific pieces of information needed to fulfill that intent.
aws lexv2-models create-intent --bot-id abc123 --bot-version DRAFT --locale-id en_US --intent-name BookFlight --sample-utterances '[{"utterance":"I want to book a flight"}]'
Real-world example
An airline builds a Lex based chatbot that understands when a customer says they want to book a flight, then guides the conversation to collect the departure city, destination, and travel date needed to complete the booking.
Common follow-ups: How does Lex integrate with Lambda to fulfill a recognized intent?;What is the difference between an intent, an utterance, and a slot in Lex?
Lambda & Serverless;Amazon API Gateway
What is Amazon Comprehend and what kinds of insights can it extract from text?
Intermediate
Amazon Comprehend is a natural language processing service that uses machine learning to extract insights from text, including identifying the overall sentiment of a piece of text, recognizing key phrases and named entities such as people or organizations, detecting the dominant language, and even identifying custom entities or classifications specific to your own business when trained with your own labeled examples.
aws comprehend detect-sentiment --text 'The customer service was excellent' --language-code en
Real-world example
A customer support platform runs every incoming support ticket through Comprehend to automatically detect sentiment, allowing frustrated or negative tickets to be prioritized and routed to senior support agents immediately.
Common follow-ups: How does Comprehend Custom Classification differ from the standard built in analysis features?;What languages does Comprehend support for sentiment analysis?
Amazon SageMaker & Machine Learning on AWS;AWS Glue & ETL
How can these AI services, such as Rekognition and Comprehend, be combined into a single automated workflow?
Intermediate
These AI services are commonly combined using event driven architectures, such as triggering a Lambda function whenever a new file is uploaded to S3, which then calls Rekognition to analyze an image or Comprehend to analyze a document's text, and the extracted insights can then be stored in a database, trigger further workflow steps through Step Functions, or generate notifications through SNS, all without requiring any dedicated machine learning infrastructure.
// Lambda function triggered by S3 upload
// calls Rekognition then stores results in DynamoDB
response = rekognition.detect_labels(image_s3_object)
dynamodb.put_item(response)
Real-world example
An insurance company automatically analyzes photos submitted with a claim using Rekognition to detect vehicle damage, then uses Comprehend to analyze the accompanying written description for sentiment and key details, combining both results to help prioritize urgent claims.
Common follow-ups: What AWS services are commonly used to orchestrate a multi step AI analysis pipeline?;How do you handle errors gracefully within an automated AI processing workflow?
AWS Step Functions;Lambda & Serverless
How do you train a custom model in Amazon Rekognition Custom Labels or Comprehend Custom Classification for domain specific needs?
Advanced
Both Rekognition Custom Labels and Comprehend Custom Classification let you provide your own labeled training examples, such as images tagged with specific product defect types or text documents tagged with custom business categories, and the service automatically handles the underlying machine learning training process, producing a custom model you can then use through the same simple API pattern as the built in features, without requiring deep machine learning expertise.
aws rekognition create-project-version --project-arn arn:aws:rekognition:us-east-1:123456789012:project/defect-detection --version-name v1 --output-config S3Bucket=my-bucket
Real-world example
A manufacturing company trains a Rekognition Custom Labels model using thousands of labeled photos of both defective and acceptable parts from its own assembly line, achieving accurate defect detection specific to its unique products that a generic model could never provide.
Common follow-ups: How many labeled examples are typically needed to train an effective custom model?;How does the cost of custom models compare to using the built in pretrained features?
Amazon SageMaker & Machine Learning on AWS;AWS Cost Management & Billing
What security and privacy considerations should be addressed when using AI services that process sensitive text, images, or voice data?
Advanced
When using services like Rekognition, Comprehend, or Lex to process sensitive data such as customer faces, personal conversations, or confidential documents, it is important to encrypt data both in transit and at rest, apply strict IAM policies limiting which roles can invoke these services, consider whether data should be processed within a VPC using interface endpoints to avoid traversing the public internet, and review each service's data retention and usage policies to ensure compliance with relevant privacy regulations.
aws comprehend detect-pii-entities --text 'Customer John Smith called about invoice 12345' --language-code en
Real-world example
A healthcare provider uses Comprehend Medical alongside strict IAM policies and VPC endpoints to detect and redact personally identifiable information from patient notes before that data is shared with any downstream analytics system.
Common follow-ups: What is the difference between Comprehend and Comprehend Medical for handling sensitive healthcare text?;How do VPC endpoints reduce data exposure risk when calling AI services?
IAM;VPC & Networking