At Micron Technology Inc, I develop reproducible Python workflows to validate machine learning models for NAND flash defect identification. I also built PyTorch inference systems to optimize DRAM circuit layouts.
I engineered Apache Airflow and SQL pipelines for sensor data, supporting real-time inference for fabrication equipment monitoring. My work also includes predictive maintenance and anomaly detection for manufacturing data.
At Hippo Insurance, I developed PyTorch and OpenCV computer vision models to identify property risks from aerial imagery. I also built NLP pipelines to extract information from insurance claim documents.
I implemented retrieval-augmented generation for policy coverage questions and developed models for smart home risk scoring, customer churn, and insurance claim anomalies. I’ve deployed models on AWS SageMaker and monitored them with Prometheus and Grafana.
