Kapil strengthened the testing framework for apache/incubator-gluten by enhancing GlutenLogQuerySuite to query Spark logs for exceptions using SQL, enabling more precise error detection during test runs. Working primarily in Scala and leveraging Spark’s data processing capabilities, Kapil focused on improving test reliability and maintainability. The work included removing redundant aggregate hash map test suites, updating DataFrame aggregation tests for compatibility, and adding documentation to clarify the changes. These updates improved test coverage for error scenarios, stabilized continuous integration, and reduced maintenance overhead, resulting in faster feedback cycles and more reliable releases. The engineering work demonstrated thoughtful, targeted improvements.