Data Scientist

Data Science & Research Indianapolis, Indiana, United States

Profile
Contacts

Summary

Data Science Intern at Blue Cross Blue Shield of Michigan
Developed an end-to-end predictive model for resource allocation using state-of-the-art machine learning on a large structured dataset with millions of records and 1,197 features.
Wrote SQL queries for data extraction, utilized PySpark for data cleaning, conducted exploratory data analysis and statistical analysis, performed feature selection and generated new features, reducing the number of features to 650.
Experimented with random forests for baseline modeling.
Built and trained a neural network model using monte-carlo dropouts to predict distributions, wrote the code to automate hyperparameter tuning. Mean absolute error was reduced by 19% with respect to baseline.
Created visualizations for stakeholders to communicate actionable insights and wrote a report about the project.