Fengtian enhanced the FlightRecord API within the pytorch/pytorch repository by making its interface publicly accessible, focusing on improving API design and external integration. Using C++ and leveraging the TORCH_API macro, Fengtian exposed multiple methods to enable broader usage by downstream tools and performance-monitoring workflows. This work increased the maintainability and interoperability of the FlightRecord component, allowing external developers to integrate with PyTorch more effectively. Fengtian also documented the changes and laid the foundation for future testing and cross-version compatibility, ensuring a stable and accessible public API surface. The work demonstrated depth in software architecture and thoughtful API evolution.