At Mastercard, I architect cloud-scale pipelines processing multi-terabyte datasets daily for AI and advanced analytics. I build distributed Spark and PySpark workflows on AWS EMR and Databricks, real-time Kafka and Kinesis ingestion, and MLflow deployment pipelines.
I've improved batch processing performance by 38%, accelerated experiment-to-production cycles by 30%, increased model prediction accuracy by 18%, and reduced data incident response time by 40%. I partner with data scientists and product teams to operationalize NLP and time-series models for fraud detection and customer insights.
Previously, I built ETL frameworks and scalable ingestion pipelines at Transol Systems and supported AI-ready pipelines at Edward Jones using AWS Glue, Redshift, S3, PySpark, Scikit-learn, and TensorFlow.
