Experience

• Main activities included learning, data pre-processing, model development, model training and testing, attending progress meetings and communicating results

• Developed an XGBoost-based model for the prediction of septic shock onset in patients with a test AUC of 0.83

• Pre-processed and used patient data from MIMIC-III and Seoul St. Mary's Hospital, which consisted of their tabular vital sign data over time

• Improved the model's AUROC (+0.04) and accuracy (4%) by feature engineering

• Devised and translated SDS and SRS documents on a product submitted for FDA 510(k) approval