At Goldman Sachs, I optimized a trade-processing pipeline with Kafka, AWS, and DynamoDB, clearing a 200,000-message backlog and reducing processing time from 24+ hours to 4 hours.
I also migrated journal processing from a legacy Union Journal architecture to a Variance Journal table using Java Spring, Spark, Snowflake, and AWS Glue. The work reduced calculation runtime by 30%, dashboard query latency by 75%, and storage overhead by 38 million records per day.
Earlier at Goldman Sachs, I built CI/CD validation controls and schema versioning for downstream machine-learning services, and integrated fraud-detection algorithms into a Java Spring Boot backend.

