Pinned Loading

1. Optimizing-Air-Travel Optimizing-Air-Travel Public

   Developed a flight delay prediction system using ML models (Random Forest, XGBoost) on historical flight and weather data, improving operational planning and delay mitigation.

   Jupyter Notebook

2. CreditGuard CreditGuard Public

   Developed a credit card fraud detection system using ML models (Logistic Regression, XGBoost) with SMOTE-based oversampling, improving fraud identification accuracy on imbalanced datasets.

   Jupyter Notebook

3. GalaxyML GalaxyML Public

   Built a deep learning pipeline using CNNs to classify galaxy morphologies from astronomical images, enabling automated analysis for astrophysical research.

   Jupyter Notebook

4. Knowledge-Distillation-Safety Knowledge-Distillation-Safety Public

   Developed a knowledge distillation framework to train compact student models using soft-label supervision, evaluating adversarial robustness (FGSM, PGD), calibration (ECE), and performance under di…

   Python

5. Docuquery Docuquery Public

   Developed a retrieval-augmented question answering system using FAISS, sentence transformers, and LLMs to generate context-aware answers from unstructured documents.

   Python

6. Metagen Metagen Public

   Automated metadata generation for unstructured documents using NLP (spaCy, Transformers, topic modeling), improving search relevance and retrieval efficiency via scalable Python pipelines.

   Jupyter Notebook