Developed a data secrecy-aware operation balancing feature for the google/heir repository, focusing on secure and efficient execution of data processing tasks. Leveraging C++ and MLIR, the work integrated a SecretnessAnalysis to dynamically determine whether operations should execute on public or secret data, optimizing the execution flow accordingly. This approach enabled security-conscious scheduling and supported compliance with data handling policies by reducing the risk footprint associated with sensitive data. The implementation demonstrated proficiency in algorithm optimization and backend engineering, laying a technical foundation for future enhancements in secure data processing and execution-flow balancing within the project’s architecture.