My passion is building intelligent systems that solve real-world problems.
Check out my projects below.
Projects
Automated Phytoplankton Classification (CAPCS)
An end-to-end computer vision pipeline for automated phytoplankton identification from ocean imagery, built at the Lamont-Doherty Earth Observatory at Columbia University. Structured 12,315 microscopy images across 10 phytoplankton classes, performed diagnostics via confusion matrix analysis, and evaluated per-class precision, recall, and F1 scores to guide dataset refinement.
Bone Fracture Detection from Medical Imaging
A machine learning system for automated detection of bone fractures from medical imaging data (X-rays). Developed and evaluated classification models to assist radiologists in identifying fractures, leveraging image preprocessing, feature extraction, and deep learning techniques for robust diagnostic support.
Loan Approval Prediction
Research and development of a loan approval prediction model at Tata Consultancy Services. Designed a comprehensive data cleaning workflow leveraging Z-Score Normalization and IQR outlier elimination on a 40,000+ entry database, resulting in 89% accuracy using Gradient Boosting with an 8% accuracy improvement and 9% runtime reduction.