Automated loan approval and disapproval decisions by processing potential borrowers banking information and historical user data reducing manual processing timeDesigned and implemented an in-house facial recognition system to verify borrowers identity by comparing a user uploaded selfie image and photos of their ID using OpenCVIntegrated a banking data providers API into our code base and built a dashboard to present the data to our internal users from each of their endpoints

• Developed an end-to-end data analysis pipeline using pandas/NumPy/sklearn that models critical metrics from large time-series datasets and is used to determine if new samples meet required specs.• Packaged the data analysis pipeline into a cross-platform GUI using PyQt5 that allows non-technical employees to run the pipeline with ease reducing data analysis time by 10 hours/week• A results table is presented to the user via QTableView and is saved on disk to a read-only .CSV file• Setup Admin and User accounts on the company computer with strict permissions to comply with CFR21: Section 11 that prevents any results generated by the data analysis pipeline from being tampered with

I create end-to-end data processing pipelines for high throughput sequencing data. Additionally, I assist in grant writing and provide mentorship to undergraduate thesis students.

Designed 2 database's using REDCap for phase I clinical trials.Developed python scripts to transform and analyze clinical datasets from REDCap.Automated data analysis pipeline for Chemiluminescence Nitric Oxide Analyzer with Python.Trained staff how to use data capture system in the clinic on an iPad.Performed experiments using an antimicrobial drug against various bacteria in the lab.Communicated results to colleagues.

Developed convolutional neural networks for deep image clustering using pytorchImplemented state of the art neural network architectures from scientific papersBecame familiar with the Google's Cloud Computing Infrastructure to utilize GPUs to train neural networks on very large datasets.