This presentation is a 10-minute lightning talk for Episode 2 of AWSUGTW's CDK SQUAD series. I, still having a role of a data architect, crafted this deck for the session. In it, I provide an overview of L1, L2, and L3 constructs, offering insights and recommendations in both layman's terms and technical jargon based on my professional experience. The highlight is a hands-on example of creating a rough CDK pattern, specifically focusing on DDB streaming into a data lake. To conclude, I introduce a GitHub repository and discuss the incremental export of DDB to S3.