Introduction to AWS Machine Learning for IoT

Explore Amazon Polly, a cloud machine-learning service that converts text to human-like speech with standard and neural voices, and review its pricing, use cases, and Raspberry Pi 4 demonstrations.

Discover how Amazon Polly converts text to lifelike speech using SSML, with multiple languages and voices and outputs in MP3, Vorbis, or PCM, and learn AWS console and CLI setup.

Demonstrates how to use Amazon Polly with Raspberry Pi 4 to synthesize text into speech, install AWS SDK and CLI, configure credentials, and play audio via Pygame.

Discover how Amazon Transcribe uses deep learning to convert live or recorded audio into text with automatic punctuation, timestamps, speaker labeling, and custom vocabulary for scalable, cost-effective transcription.

Learn how Amazon Transcribe converts speech to text, with batch and streaming options, using the AWS management console or CLI, including S3 workflows and real-time transcription.

Demonstrate setting up Amazon Transcribe on a Raspberry Pi 4 by recording audio, uploading to S3, and starting a transcription job with AWS CLI and boto3.

Explore Amazon Rekognition’s deep learning image and video analysis, identifying objects, text, people, activities, scenes, and inappropriate content, with face detection, facial analysis, and face comparison.

Explore how Amazon Rekognition processes images and videos through image and video APIs, covering labels, faces, text detection, and storage vs non-storage operations in S3 and collections.

Get started with Amazon Rekognition using AWS CLI, AWS SDK, or the Management Console to perform image and video analysis, including label detection, face analysis, and face comparisons.

Demonstrates configuring AWS CLI and IAM permissions, uploading reference images to S3, indexing faces with boto3 on a Raspberry Pi 4, and performing real-time Rekognition face matching.

Demonstrate Amazon Rekognition on a Raspberry Pi 4 with a 5 MP camera module, covering image capture and matching faces via Rekognition's search by image against indexed faces.

Create an end-to-end proximity interactor for your door on a Raspberry Pi 4 using ultrasonic sensing, microphone input, camera, speaker, and buzzer, with Amazon Polly, Transcribe, and Rekognition.

Set up an IAM user and an S3 bucket, configure AWS via the AWS CLI, and explore the proximity human interactor project using Rekognition for face indexing.

Demonstrates an end-to-end proximity identity system using audio recording, Amazon Transcribe, S3 storage, Rekognition face matching, and an ultrasonic sensor.

Explore Amazon SageMaker as a fully managed machine learning service that unifies data labeling, model building, training, deployment, and monitoring on AWS, with an architecture view and workflow.

Explore SageMaker training options, including built-in algorithms and custom code with TensorFlow or MXNet, plus Spark integration. Review SageMaker pricing with free tier and pay-as-you-go models.

Explore the Amazon SageMaker dashboard and Studio, including Canvas and Ground Truth, to learn the end-to-end workflow from data preparation to deploying endpoints with edge manager.

Upload your dataset to S3, create an Amazon SageMaker Canvas model, train a simple Celsius-to-Kelvin conversion, and generate predictions without coding.

Explore SageMaker Edge Manager for deploying, monitoring, and updating ML models on a fleet of edge devices, including deployment options, security, and real-time latency insights.

Explore an end-to-end AWS SageMaker edge manager workflow, covering data visualization, model building, packaging, deploying on edge devices, and resource cleanup with Greengrass v2 or IoT jobs.

Recap the beginner-friendly overview of AWS machine learning for IoT, highlighting Polly, Rekognition, Transcribe, and SageMaker, with Raspberry Pi 4 demos and next-steps for IoT projects.