Mastering AI-Driven Enterprise Apps with AWS & AI

Develop intelligent data applications by prototyping machine learning models, refining feature sets and preprocessing, benchmarking performance, and deploying models that run on schedule with ingestion and storage using AWS.

Explore key AWS data technologies, including S3, Redshift, and data pipelines, and learn how to stage, ingest, transform, and store data for scalable machine learning architectures.

Understand how each chapter's project directories—ingestion, destination, pipeline, model, and stream—are organized in S3, and how to use them with AWS services like Kinesis, Data Pipeline, and machine learning workflows.

Explore project workflow for intelligent data applications. Define a real-world problem, design an architecture, deploy and monitor scalable AI pipelines across storage and staging.

Automate selective data transfer from a staging area to a destination using JSON-defined data pipelines, S3 storage, and on-demand EC2 instances for scalable ETL-like extraction through partitioning.

Load ingestion data into the S3 ingestion directory, import the JSON pipeline definition, and activate an ETL pipeline that copies telemetry files within S3, with logging and IAM setup.

Finish an etl pipeline in AWS data pipelines using cli, gui, or json definitions. Configure s3 sources and destinations, on-demand ec2 resources, alarms, and activation to monitor execution.

Explore a modular, scalable ETL workflow on AWS that migrates data between S3 buckets, runs automated pipelines, and lays the groundwork for machine learning in enterprise apps.

Design an automated machine learning app on AWS by loading data from S3 via pipelines, running a Python script on EC2, and saving scores and logs to S3.

Deploy an end-to-end machine learning pipeline on AWS, load Iris data from S3, run a Python script on a small Linux instance, and configure logs, IAM roles, and pipeline definitions.

Explore how to define and run a machine learning pipeline on AWS using a JSON pipeline definition and an EC2 script, training multiple classifiers and exporting scores to S3.

Review the deployed ETL and machine learning workflow using Amazon data pipeline and EMR, with data preprocessing, a custom image and Python scripts for scalable offline training and deployment.

Extend the architecture with a database layer using RDS and Redshift, loading S3 data and model outputs into database tables to support personalized user experiences, monitoring, and reporting.

Build automated etl pipelines to schedule and load data from S3 into Redshift. Use csv or json formats, configure copy options with staging for transformation, and apply upsert or truncate.

Explore loading data into Redshift with JSON pipeline definitions, including pre-processing and transformations, plus CSV options. Learn to handle undefined values and nulls, diagnose STL load errors, and tune schedules.

Load data from S3 into a MySQL database using a JSON-defined AWS data pipeline, create tables, and insert data with SQL and copy activities.

Load data from S3 into RDS or MySQL DB using a pipeline, compare with Redshift, and review the pipeline definition, instance timing, and JSON parameters.

Integrate a Redshift and Redis data layer to accelerate model iteration, enabling machine learning outputs to feed database tables, with cost and tool considerations as AWS shifts to alternatives.

Design a streaming data ingestion pipeline that uses Kinesis to bring telemetry into S3, performs real-time analytics and quality checks, and loads data into Redshift for downstream ML and applications.

Configure the kinesis data generator with CDBG to test kinesis streams and firehose using a template. Create a Cognito user and stream, log in, set region, and start sending data.

Configure the Kinesis data generator to create telemetry with randomized user IDs, names, dates, and weighted values, then test sending records to Kinesis streams and S3.

Build a kinesis streaming analytics workflow by setting up a kinesis stream, testing with generated data for outlier detection, and wiring a firehose or stream endpoint.

Learn to correct Kinesis Analytics schema by manually setting uppercase column names, adjusting row and column limits, and resolving hyphen delimiter and data-type mismatches before building real-time queries.

Build anomaly detection pipelines with Kinesis streaming analytics using random cut forest to score anomalies. Configure streams, pumps, apply tumbling and sliding windows, and push results to S3 via firehose.

Integrate Kinesis Firehose, streams, and analytics to enable real-time analytics and ingestion-time processing of live data. Assess cost and use cases for anomaly detection and aggregations within end-to-end data stack.

Discover Amazon machine learning: quickly build and deploy models with a ready API endpoint, using data from S3 or Redshift, and monitor performance with CloudWatch.

Prepare data sources from s3 csv datasets for Amazon Machine Learning using iris, air quality, and diabetic retinopathy data to train multiclass, regression, and binary models, using benzene as target.

Select and customize the iris data set in Amazon Machine Learning, adjust the recipe and random evaluation splits, and apply normalization to build and evaluate regression and classification models.

Evaluate and compare Amazon Machine Learning models from iris to diabetic retinopathy, review evaluation metrics, thresholds, and confusion matrices, and deploy an endpoint for real-time predictions.

deploy a model as a real-time endpoint, call it with Python and boto3, and retrieve live predictions while reviewing model details and endpoint information.

Deploy real-time and batch Amazon Machine Learning endpoints to predict new data, evaluate models, and automate preprocessing with S3 integration, enabling accessible ML for enterprise apps.

Learn to design end-to-end, automated AI solutions on AWS by chaining Kinesis streams and analytics with data pipelines, automated preprocessing, and deployed ML endpoints.