Worked on stabilizing model configuration in the Future-House/paper-qa repository by enforcing strict temperature constraints for OpenAI o1-series models. Addressed a configuration safety issue by implementing a Pydantic settings-level validator to ensure the temperature parameter is always set to one for o1-preview and o1-mini models, reducing the risk of inconsistent or unintended model outputs. Enhanced reliability by adding automated tests to verify the new validation logic, improving test coverage and deployment consistency. Utilized Python for backend development and configuration management, focusing on robust validation and maintainability. The work prioritized safe, predictable model behavior across different deployment environments.