Anshul Saha developed version-specific tests for LLM semantic convention attributes in the pydantic/logfire repository, focusing on integrations with Anthropic and OpenAI. By implementing automated test suites in Python, Anshul validated attribute handling and data formats across providers, ensuring reliable message conversion and response processing. The work emphasized API integration and cross-provider compatibility, addressing the need for robust deployment safety in LLM features. Although no major bugs were fixed during this period, the contribution deepened the test coverage and improved reliability for future releases. Collaboration with other contributors further strengthened the technical depth and maintainability of the codebase.