Developed and delivered a WandB logging enhancement for the apple/axlearn repository, focusing on flattening nested dictionaries to improve experiment log clarity and dashboard usability. The solution standardized key formats within logs, enabling more effective filtering and consistent data presentation across experiments. This work leveraged Python for implementation, with attention to data logging best practices and unit testing to ensure reliability. By addressing the challenge of nested dictionary structures in experiment tracking, the enhancement facilitated better observability and streamlined debugging, supporting data-driven decision-making and smoother collaboration for teams working with the axlearn codebase. No major bugs were addressed during this period.