Developed an efficient Parquet file reading feature for the aws/aws-sdk-pandas repository, focusing on processing large datasets within memory-constrained environments. The solution introduced chunked reading per row group, leveraging Python for data processing and memory optimization. By reading Parquet files in smaller, manageable chunks, the implementation reduced peak memory usage and improved throughput, enabling the handling of larger workloads without exceeding system limits. This work established a foundation for future enhancements such as streaming and partial reads. The approach demonstrated a strong understanding of scalable data processing, with careful attention to performance and resource management using Python and data engineering techniques.