• Synthetic data team that used CTGANs on real datasets to produce synthetic data for customers throughout the organization. Used for data anonymization and to safeguard customer privacy.• Used containers within Microsoft Azure to create data warehouses and used Azure ML Studio to create notebooks, load and save datasets, run experiments and store models. Utilized GPUs for faster run time. • Provide clients high-level overviews of the models’ results by explaining Gaussian Mixture Log Likelihood, Chi-Squared, Logistic Regression Detection and Discrete Kullback-Leibler Divergence.

• Used the ELK stack and Dynatrace to capture user data instances to train and test a neural network that recommended drug names and drug dosages to customers.• Led a team of Microsoft consultants that implemented natural language processing techniques to improve model accuracy by 50%.• Created ETL pipelines that deployed Jenkins jobs which ran Dockerfiles and Python files to pull data from an Elasticsearch index and cleansed and built datasets.• Managed a Github repo for an ETL pipeline that allowed the team to submit pipelines locally to Microsoft Azure that registered and trained models and deployed APIs.• Wrote Python functions that sent post requests to check HTTP status codes. Used multithreading to send json objects to various endpoints and improve speed. Analyzed the responses using Python and Microsoft Azure.• Compared multiple endpoints’ drug name recommendations. Wrote Python scripts that transformed the data and compared 1000s of drug lists’ responses and confirmed each respective drug list returned the same recommendation.• Calculated the hit rate for each model by writing Python code which verified if the drug recommended by the model was the drug removed from the original drug list.