Worked on the facebookresearch/param repository to address a critical bug in execution trace mapping for multi-process and multi-threaded workloads. Focused on improving the accuracy of process ID to thread ID and node ID associations, the developer updated the tracking logic to utilize PyTorch profiler thread annotations, moving away from deprecated root thread identifiers. This Python-based solution enhanced the reliability of trace analysis and ensured consistent index mapping across runs, which is essential for effective performance debugging. Leveraging skills in code analysis and debugging, the work laid a foundation for more robust trace analytics in future profiling and performance workflows.