Start here in AWS
The most useful guides in this topic, not just the most recent.
AWS Guides & Tutorials
AWS for data engineers — S3, EC2, Lambda, Glue, Redshift, Athena, EMR and the certification roadmap that gets you hired.
AWS articles
S3 Lifecycle Policies and Storage Cost Savings
S3 Lifecycle rules move objects to cheaper storage classes and delete them on a schedule. This guide explains transition and expiration actions, builds a three-rule configuration for a data lake bucket with boto3, and sets out the charges that catch people out: small objects, minimum durations, versions and archives.
AWS Glue vs EMR vs Lambda for ETL
Glue, EMR and Lambda can each run ETL code on AWS, and they are built for very different sizes of job. This comparison sets out the hard limits of each, how data volume, control and operations decide the choice, how each is billed, and why most real pipelines end up using two or three of them together.
Building a Serverless Analytics Stack on AWS
A serverless analytics stack on AWS has no cluster to run: S3 for storage, Glue for the catalog and transformation, Athena for SQL, Lambda and Step Functions for triggers and orchestration, and a BI tool on top. This guide builds it layer by layer, lists what each service bills for, and says where it stops fitting.
Building an ETL Pipeline on AWS Step by Step
A step-by-step ETL pipeline on AWS, run end to end: extract raw files from S3, validate 2.96 million real rows with named rules, load cleaned Parquet into a partition, register it in the Glue Data Catalog, write a reconciliation report, and rerun without duplicates. Each step is mapped to its production service.
Event-Driven Pipelines with Lambda and S3
An event-driven pipeline starts work the moment a file arrives in S3. This guide explains S3 event notifications, writes a Lambda handler that validates and converts a real file, and runs it twice on the same event to show idempotency. It also covers encoded keys, loops, failure handling, and when to add a queue.
Data Modeling in Redshift: A Practical Guide
How to model data in Amazon Redshift: a star schema written as Redshift DDL, why primary and foreign keys are declared but not enforced, how to choose column types, where distribution and sort keys fit, what to do with semi-structured data, and when a materialized view or a wide table is the better answer.