Julian Benny

Grammar Correction Tool for Books and Journals: Instrumental in the development of a data pipeline for extracting raw textual data from .docx and .xml files. Leveraged Transformer-based models to elevate text classification and grammar precision, with corrections presented as track changes for seamless review and editing, greatly enhancing copy editing efficiency. Alt Text Generator: Led the collaborative effort to develop an end-to-end alt text generator for docx manuscripts, significantly advancing content accessibility. Adaptation and fine-tuning open-source LLMs for optimized generation of descriptive alt texts. Correcting Raw Author Biographies: Collaborated in automating the correction process for sequences of events in author biographies. Contributed to fine-tuning pre-trained Transformer models, significantly improving the editing and polishing of raw biographies.

Spearheaded the web automation module for a major project, utilizing Selenium, BDD methodologies, and Java, resulting in enhanced testing efficiency and reliability. Implemented API automation with Python and REST, reducing manual effort and increasing coverage. Central to the migration of the testing suite to a BDD framework with Python, improving maintainability and agility.