Advanced Orchestration Patterns with AWS Step Functions

A deep dive into production-grade orchestration design using Step Functions. Standard vs Express workflows, Map state scaling, payload management, retry semantics, and failure isolation. A dedicated JSONata transformations exploration, real-world patterns, performance considerations, and debugging strategies.

For many teams, AWS Step Functions are primarily used to orchestrate Lambda functions. In practice, this often leads to workflows filled with “glue Lambdas” that only transform payloads or proxy HTTP calls between services.
In recent years, AWS has significantly expanded Step Functions with features such as JSONata-based transformations, native HTTP endpoint integrations, and EventBridge Connections for managed authentication.
Together, these capabilities allow workflows to transform data and call external services directly, reducing the need for intermediary code and simplifying serverless architectures.
In this talk we will explore how these features change the way Step Functions workflows are designed. Through practical examples, we will look at when to transform data directly in the workflow, when to call services without Lambda, and how to build cleaner, more maintainable serverless systems.