At Polymerize, I build machine learning pipelines for single-stage and multi-stage prediction systems, improving model accuracy by 10% and accelerating inference through ONNX conversion and optimization. I also developed a Bayesian Optimization experiment recommendation system that reduced experimentation cycles by 50%.
Previously at Corteva Agriscience, I productionized computer vision systems for seed quality classification and real-time corn inspection using YOLOv8, EfficientNet, and ByteTrack. These systems achieved up to 0.91 F1-score and reduced manual inspection effort by 50%.
My work spans clinical NLP at University of Florida Health, data engineering at Tata Consultancy Services, and healthcare research publications. I use Python, SQL, transformer models, computer vision, and cloud tools to turn complex data into practical, transparent AI products.
