max sivolella

Data Scientist

Stanford University•  October 2023 - Present•  Remote

• Led research on predicting glaucoma progression by developing a multi-modal neural network in PyTorch, surpassing the accuracy of leading research by 15%. Model projected to prevent irreversible vision loss in 700+ patients, saving $6M in surgical costs annually. • Created an ETL pipeline to parse through and load 100M+ rows, 150+ tables, and 1+ terabytes of medical data into a centralized database in Google Cloud Platform using SQL and Python scripts, establishing a statistically relevant database of patients for predictive modeling. • Implemented a parallel computing pipeline using PySpark for feature engineering of 400+ predictors, reducing runtime by 35%. • Developed a random forest model, integrated with EHR data on the residuals of the primary model, improving RMSE by 10%. • Selected to present research paper at the American Medical Informatics Association conference, with publication in its final stages. • Applied statistical modeling using functional Principal Component Analysis on sparse time series (longitudinal) data to construct a global disease progression function with scikit-learn, stats models and R, helping doctors to visualize future vision loss trajectories of patients.

Healthcare Data Analyst

911 Health•  September 2021 - December 2022•  Santa Monica

• Optimized workflow procedures at 2 large Covid testing sites (500+ patients weekly) by designing a more efficient testing schedule, resulting in a 25% reduction in testing time, saving over 15,000 hours of labor and reducing company expenses by $450,000 yearly. • Implemented a data pipeline by creating and integrating a web-app with the company database allowing employees to automatically update patient profiles, saving clients and the company $1.21M in redundant testing fees and reducing plastic waste by 4,500 lbs annually.

Product Analyst

Injectsense•  June 2020 - September 2020•  Emeryville, CA

• Wrote a 10-page technical report, including statistical analyses, visualizations and dashboards using SQL and Tableau, to analyze trends impacting the company and industry, contributing to investor meetings and helping secure $3.4M in Series B funding. • Edited documents for submission to the FDA contributing to the product earning a breakthrough device designation, increasing investment potential and expediting product development and review, reducing time to market access by 1.5 years.