Worked on enhancing observability and reliability within the rungalileo/galileo-python repository by addressing logging issues in LLM tooling integration. Focused on improving the OpenAI decorator, the work involved updating input data extraction to accurately capture tool definitions and ensuring this information was correctly passed into the LLM span for comprehensive logging. Deprecated aliases for tool-related parameters were removed to prevent mislogging and streamline telemetry. The solution leveraged Python, API integration, and the decorator pattern to resolve a persistent bug, resulting in more accurate end-to-end monitoring of LLM interactions and a simplified, more maintainable logging infrastructure for future development.