Contributed to PX4/PX4-Autopilot by delivering targeted improvements in performance monitoring and backend processing using C++ and Python. Addressed a critical bug in the performance counter reset logic, ensuring that internal statistics such as mean and M2 are fully reset to maintain accurate telemetry and analytics data. Later, implemented a feature upgrade in the px_uploader tool by replacing a manual CRC32 calculation with Python’s zlib.crc32, resulting in a substantial speedup for large firmware uploads. These changes enhanced the reliability of performance metrics and streamlined deployment workflows, demonstrating a methodical approach to embedded systems and backend development challenges.