- lead RAG developer for internal generative AI chatbot, including document ingestion/transformation, chunking strategies, dynamic prompting, conversation flow state- lead end-to-end development of supervised classification models for fraud detection

- lead the end-to-end development of supervised classification models - understand business requirements - collect and transform large datasets on Hadoop Distributed File System (HDFS) with Apache Impala (SQL) - build pipelines to feed data to MLlib supervised classification models with Python, PySpark, Scala, and Spark SQL - improve model performance by applying various feature selection techniques (forward/backward stepwise selection, PCA), imbalanced data techniques (over/undersampling, SMOTE), hyperparameter tuning with grid search algorithms - deploy final model to production, including planning and scheduling batch prediction jobs, and working with admins to move the model to production environments- regularly communicate with the client in non-technical jargon to explain the functionality of the model and receive input to make any necessary changes- manage Git repository to ensure concise, quality, and readable code to avoid technical debt

- using Apache Spark to build supervised machine learning models (random forest, GBT, logistic regression) to improve the legacy overpayment detection process- communicating regularly with the Director's Office about how the machine learning models improve the business- manipulating millions of data utilizing SQL and Apache Spark- led development of successful proof-of-concept unsupervised machine learning models to identify problematic healthcare claims

Tutored Virginia Tech students in various calculus-based courses.

Work involved showing examples of leadership to children ages 8-11 through day to day field trips and recreation, as well as heavy collaboration and communication with coworkers and parents.

Parsed through physical data entries, cleaned, and transferred to database in Eclipse.