Worked on enhancing deployment configurability and observability for Flink workloads on AWS EKS within the zipline-ai/chronon repository. Developed an environment variable-driven system allowing dynamic configuration of Kubernetes pod labels for Flink JobManager and TaskManager, supporting flexible metadata attribution and improved traceability. Implemented a robust Scala utility to parse comma-separated key-value pairs, handling edge cases to ensure consistent PodTemplateLabels across components. Comprehensive unit tests were written to validate parsing logic and maintain CI stability. The work aligned with existing FLINK_EKS conventions, leveraging skills in Kubernetes, AWS EKS, and backend development to enable future enhancements in metadata management.