Gangadhar J

Ai Ml Engineer @PNC

Designed, trained, and deployed advanced fraud-detection models using TensorFlow and AWS SageMaker, improving transaction validation accuracy by 20% and reducing false positives, which directly strengthened customer trust and compliance outcomes. • Engineered scalable data pipelines with Apache Spark, AWS Glue, and S3 to process millions of daily transactions for realtime model scoring and anomaly monitoring, ensuring consistent, high-quality data ingestion. • Implemented end-to-end MLOps frameworks with Docker, Kubernetes, GitLab CI, and MLflow to automate retraining, testing, and deployment workflows, cutting model delivery cycles by 40% and improving system reproducibility. • Built and maintained model monitoring dashboards in Power BI and CloudWatch to visualize drift, anomalies, and inference metrics; automated alerting and retraining triggers reduced downtime by 25%. • Partnered closely with data engineers and product teams to integrate models into legacy banking systems through REST APIs and batch pipelines, ensuring seamless adoption and zero-downtime deployments. • Conducted A/B testing and model performance benchmarking to evaluate new ML algorithms and hyperparameter settings, improving predictive accuracy and model stability across multiple financial use cases.