Shanmukha Sai Jayavarapu
Machine Learning Engineer @KPMG US
Signup · Get unlimited contacts
WORK HISTORY
Machine Learning Engineer @KPMG US
Project: NextGen Financial AI Risk & Fraud Platform· Engineered real-time fraud detection pipelines using Kafka, PyTorch and Graph ML, scaling to process millions of transactions per hour with sub-second latency.· Reduced fraudulent transaction losses by 38% by deploying advanced Graph Neural Networks (PyTorch Geometric) and anomaly detection models for suspicious transactions and network analysis.· Built and productionized credit risk scoring models leveraging XGBoost, LightGBM and CatBoost, achieving a 22% uplift in predictive accuracy and driving more accurate loan approval decisions.· Applied Explainable AI frameworks (SHAP, Captum) to ensure compliance with Basel IV and NIST AI RMF, delivering transparent and regulator-approved decision pipelines.· Established CI/CD pipelines for ML models with MLflow, Kubeflow, Jenkins and ArgoCD, ensuring seamless deployment, automated retraining and uninterrupted fraud detection operations with 99.97% uptime.· Designed regulatory-compliant data lakehouse architecture using Delta Lake and Databricks, enabling secure, auditable and large-scale financial data processing for fraud and risk analytics.· Monitored and mitigated model drift & bias with Arize AI and WhyLabs, deploying auto-alerts that reduced fraud detection downtime by 60% while maintaining fairness and reliability in financial models.· Delivered real-time executive risk dashboards in Azure Synapse Analytics, providing business units with unified insights into fraud detection, credit risk and regulatory compliance.
EDUCATION
University of Central Missouri
Master's degree
Lakireddy Bali Reddy College of Engineering
Bachelor of Technology - BTech
ABOUT SHANMUKHA SAI JAYAVARAPU
I create intelligence systems that work in situations where the stakes are high and the…
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.