Wil Ortiz developed logprobs support for Azure OpenAI GPT-5.2 within the BerriAI/litellm repository, focusing on enhancing observability and analytics for Azure deployments. He implemented the logprobs parameter specifically for the GPT-5.2 model, refactoring the codebase to ensure model-specific reliability and maintainability. Using Python, Wil applied backend development and API integration skills to address an Azure OpenAI logprobs handling issue, improving the robustness of the integration. He also expanded the test suite to verify correctness and coverage, demonstrating a disciplined engineering approach. The work delivered targeted business value by enabling token-level probability insights for enterprise users.