Developed a backpressure-enabled Kafka Connect Adapter for the datastax/pulsar repository, focusing on improving reliability and throughput during high write loads. Leveraging Java and Kafka, introduced batch-based write control with a locking mechanism to enhance concurrency safety and streamline the write path. Refactored the writer coordination into a signalWriters method, prioritizing performance optimization over rigid batch size limits. These changes stabilized the adapter’s write operations, reduced stalls, and improved maintainability through targeted code cleanups and style refactors. The work emphasized concurrency and performance optimization, laying a foundation for future scalability and more robust data integration workflows.