Worked on the Agenta-AI/agenta repository to enhance initialization observability and test reliability within the Python SDK. Developed a feature that issues a RuntimeWarning when litellm callbacks are triggered before ag.init(), preventing silent failures and improving developer diagnostics. Strengthened unit tests using pytest to ensure that pre-initialization warning paths are clearly reported and failure modes are explicit, addressing robustness concerns highlighted in issue #3171. Focused on aligning instrumentation changes with existing patterns and documentation, the work improved CI reliability and developer guidance. Demonstrated skills in Python, SDK development, and comprehensive unit testing to support maintainable and observable code.