Data Scientist (Clinical Trial Research)
CurrentPioneering incorporation of AI/ML into clinical and commercial pharma data platforms by introducing three models:- Smart Trial Design: Random forest model fit to 150,000 clinical trial designs predicts likelihood of user-designed trial's successful completion. Model surpassed accuracy benchmark set in prominent scientific paper by 5%. Visualized model outputs using SHAP values and D3.- Intelligent Data Review: Generative AI (GAN architecture) for anomaly detection in SDTM data. Autonomously preprocessed data, built deep learning model using Keras and TensorFlow, and deployed with AWS. Demonstrated and presented end-to-end model function live to technical pharmaceutical audiences.- Claims Data Chatbot: Integrated local LLMs to enable natural language queries on insurance claims data, leveraging SQL for big data processing and ensuring protection of sensitive information. Designed frontend with React.