Worked on the apache/datafusion repository to deliver a targeted performance optimization for dynamic filter updates in aggregate functions. Focused on backend development and data processing using Rust, the approach involved conditionally updating filter predicates only when the min or max bounds changed, thereby reducing unnecessary computations during query execution. This change addressed dynamic filter update overhead and resulted in an average throughput improvement of around 20 percent when dynamic filters were enabled, as confirmed by benchmarking. All modifications were validated against existing test suites and regression benchmarks, ensuring stable performance without user-facing changes and maintaining alignment with tracked project issues.