Worked on the liguodongiot/transformers repository to enhance reproducibility in distributed training workflows. Developed a feature that updates the seed_worker function to derive random seeds from both worker_id and rank, ensuring consistent results across multiple worker processes. Modified the DataLoader initialization to integrate this new seeding logic, which improves the reliability and consistency of distributed experiments. Added targeted test coverage to validate correct seed propagation and reproducibility in multi-worker scenarios. Utilized Python and PyTorch, with a focus on distributed computing and data processing. This work streamlines debugging and iteration, contributing to more dependable and efficient machine learning pipelines.