At the University of North Texas, I worked on Physics-Informed Neural Networks in PyTorch for environmental forecasting, improving long-horizon accuracy by 18% over LSTM, SARIMAX, and Linear Regression baselines. I also built ETL pipelines on AWS Glue and S3 and deployed Streamlit dashboards for model performance monitoring.
At Cargo Exchange India Private Limited, I built Python ETL pipelines for the Freight Index System, processing 50K+ daily freight records and cutting manual reporting time by about 40%. My projects include a career matching platform with a weighted Jaccard Similarity service and a multi-agent system for automated ML.

