Worked on the snowflakedb/snowpark-python repository to refactor the time_series_agg function, introducing RANGE BETWEEN INTERVAL windowing to enhance both performance and accuracy for time series aggregations. The technical approach involved deprecating and removing the sliding_interval parameter from the core logic, which simplified ongoing maintenance. Comprehensive tests were added to cover a variety of window sizes and sub-day intervals, ensuring correctness and preventing regressions. This work leveraged Python, SQL, and Snowpark, and aligned with the goal of optimizing time series analysis workflows. The changes resulted in measurable performance gains, reduced technical debt, and improved test coverage for future development.