Yang Jenny Song

Created an enterprise healthcare data platform on Azure, consolidating data from AWS and GCP into a unified analytics environment supporting 20k+ records. Developed and maintained NLP and ML systems using Python, PyTorch, and Hugging Face (T5, BERT, SBERT), analyzing over 5 million clinical and operational text records. Built data pipelines using Python, SQL, and Hadoop, reducing processing time by 35% and accelerating ML experimentation cycles. Automated end-to-end deployment of four critical ML pipelines using Terraform and IAM, increasing deployment frequency by 30% and reducing infrastructure costs by 15% across environments. Delivered AI-enabled features for telehealth platforms, supporting video consultations, secure messaging, and real-time patient data ingestion within Agile teams.

Drove the application of innovative ML techniques to structured datasets, accelerating data processing time by 45% across 15+ multinational clinical research projects. Led the development of NLP and ML solutions using Python, TensorFlow, Keras, and scikit-learn, including a chatbot that reduced response time and improved operational efficiency. Contributed to the training and evaluation of GANs and anomaly detection models, generating synthetic test data and reducing testing costs by over 20%. Collaborated with researchers and engineers on data collection, reporting, and regulatory compliance to ensure reproducible, ethical ML outputs. Executed feature engineering and model validation for predictive analytics, contributing to a 23% reduction in unplanned downtime.

Conducted exploratory and statistical analysis on high-dimensional healthcare datasets to identify patterns and risk signals supporting over 10 clinical and operational initiatives. Developed predictive and descriptive models to improve forecast accuracy and decision-making. Applied advanced feature construction on temporal and categorical data, enhancing signal quality and downstream model performance. Designed and evaluated experimental frameworks to ensure statistically sound, reproducible insights. Partnered with domain experts to translate analytical findings into actionable recommendations.