Projects

RSA-Based Private Set Intersection

Tech: Python, secrets, socket, json | Repo: GitHub Repository

• A simple Python-based implementation of the Blind RSA-based Private Set Intersection Protocol.
• It is a basic example of Secure Multi-Party Computation, ensuring two parties can find their intersecting data (modelled as sets) without exposing unshared elements.

Short-Term Temperature Prediction using Machine Learning with Explainable-AI

Tech: Python, scikit-learn, NGBoost, SHAP, LIME | Repo: GitHub Repository

• Developed supervised ML pipelines for short-term minimum and maximum temperature forecasting using multivariate meteorological data.
• Performed extensive feature engineering and evaluated six ML models, with Gradient Boosting Regressor and MLP achieving the best performance.
• Applied explainable AI (SHAP, LIME) to identify dominant climatic drivers, improving interpretability and scientific usability.
• Published in Springer (Theoretical and Applied Climatology).
• Extended the framework to probabilistic forecasting using NGBoost, jointly predicting mean and uncertainty, with Bayesian hyperparameter tuning and cross-city robustness analysis.

A Python Framework for Secret Image Sharing

Tech: Python, Galois, PIL, NumPy | Repo: GitHub Repository

• Developed a cryptographically correct Python library for (k, n) Visual Cryptography, Lossless (k, n) and (t, k, n) Shamir’s Secret Sharing on images, and Regional Secret Image Sharing.
• Ensured exact reconstruction using finite-field arithmetic (order 256).