Josna John

Background

About Josna John

After earning my engineering degree from RSET India, I moved to the United States with my family to start a new chapter. Like many first-generation immigrants, I took the first opportunity I could find to support our transition by working in project management at a commercial flooring company. It was not a traditional tech path, but it taught me how to manage complexity, lead under pressure, and deliver when the stakes are high. While I learned a great deal managing operations and blueprints, my true interest was in solving meaningful problems through data and machine learning. That conviction brought me to UC San Diego, where I recently completed my M.S. in Machine Learning and Data Science. I leaned into applied AI, particularly in healthcare. As a Machine Learning Researcher at UC San Diego Health, I worked on automating brachytherapy treatment planning using deep learning models, replacing time-intensive manual workflows with clinically viable, AI-driven ones. I also co-authored GS-TransUNet, a hybrid Vision Transformer + 2D Gaussian Splatting model that outperformed state-of-the-art benchmarks in skin lesion classification and segmentation, published in the SPIE Journal of Medical Imaging. What motivates me is building machine learning systems that not only perform in theory but also create real-world impact. Whether it is optimizing cancer treatment or helping analysts extract insights from complex documents, I care about turning research into results. If you are working on something meaningful in ML, GenAI, or healthcare, I would love to connect🤝.