In my independent research, I built an end-to-end exoplanet detection pipeline using NASA's Kepler flux dataset. I used PCA-based feature engineering and SMOTE to address class imbalance, and improved ROC-AUC from 0.50 to 0.64 with cost-weighted XGBoost.
For a pneumonia detection project, I built and regularized a CNN classifier for chest X-rays, using image augmentation and class weighting to optimize for sensitivity to the positive class.
As a B.Tech Computer Science and Engineering (Data Science) student at SRMIST, I’ve worked across machine learning, deep learning, NLP, and generative AI. My toolkit includes Python, TensorFlow, scikit-learn, and XGBoost.
