AWS for a ML Engineer
Store training data, run batch and event-driven compute, and stand up inference endpoints.
Services that matter
Amazon S3Storage
The default data lake. Training datasets, feature stores, and model artifacts live here and are read directly by training jobs and batch inference, which is why bucket layout and object-key conventions end up being a data-engineering decision rather than a storage one.
AWS LambdaCompute
Relevant to this role.
CloudWatchManagement & Governance
Relevant to this role.
Build it, graded
Host a static website on S3
Object storage, static hosting, and public read access — provisioned in CDK, not clicked through a console.
Run code without servers on Lambda
Event-driven functions, pay only for the compute time used — provisioned in CDK, not clicked through a console.