Prasanna Kothari

I'm a Computer Science and Engineering undergraduate building data-driven AI and web applications with Python, machine learning, Transformers, and modern web technologies.

I've analyzed WHO healthcare datasets and built Random Forest regression and classification models, identifying relationships between workforce density, mortality, and economic status. My predictive modeling achieved an R² of 0.63 and 71% accuracy.

I built a Transformer-based Seq2Seq model with cross-attention to generate multilingual headlines across six languages, achieving a ROUGE-1 score of 43.1 and 2.8× faster inference. I also evaluated eight activation functions in deep neural networks, where GELU achieved 93.72% accuracy in 46 epochs.

Alongside AI work, I've developed an expense-tracker web application with CRUD operations, authentication, responsive reporting, data security, and CSV/PDF export. I'm seeking opportunities to build efficient, scalable software while continuing to learn emerging technologies.