• Developed and owned a python package using Tesseract and OpenCV for OCR based data extraction from batches of documents with interchangeable content templates

• Implemented scalable parallelization of document OCR using Dask

• Configured CI/CD using Docker and Jenkins to automate the testing and deployment

• Automated a monthly assurance process where data on 10 credit card statements was manually checked for accuracy; created a new pipeline which ran on production data and allowed a statement to be processed every 2 seconds

• Lead the development of a scalable Insight Engine module leveraging Dask to analyze millions of segmentations within NLP datasets and surface subsets of metrics which stand out with statistical significance; this functionality became a core offering of the product

• Implemented features such as user permissions within the webapp including both backend API engineering using Flask and frontend design in React

• Transitioned a number of inhouse NLP models to GCP’s AI Platform for easier scaling and better versioning

• Automated miscellaneous DevOps tasks such as instance healthchecks, backup verification and new model deployment to AI Platform