Software Developer
Current- Configured a new Tesseract model (lang) using 200 base truth images to improve the overall Optical Character Recognition (OCR) accuracy by 30%. - Processed over 12,000 unique DICOM dose reports with 100% accuracy.- Participated with a development team to create an image segmentation algorithm in Python that continuously generatesbase truth files in a Json structure that analyzes images with 100% accuracy.- Directed a team of two software engineers to implement a full-stack application to upload and process patient summaries.- Utilized NGINX to create reverse proxy servers to enhance security for a full-stack application.- Migrated a legacy application from Flask to Django to implement custom fields and increase data security bycommunicating with a SQL server.- Leveraged an open source asynchronous task queue (Celery) that communicated with a message broker (RabbitMQ) to upload and receive data from AWS which decreased load times by 84%.
