This work proposes Secure FedRenyi, a privacy-preserving extension of fairness-aware federated learning that protects sensitive client fairness statistics during aggregation.

The method combines cryptographic secure aggregation using Diffie–Hellman-based pairwise masking with Laplace differential privacy for gradient updates.

DataEaze is a no-code data analytics and modeling platform that automates the complete workflow from data upload to actionable insights.

The first version was developed at CUSB under the guidance of Dr. Sandeep Kumar. The upgraded version was later developed at NIELIT Patna under the supervision of Shri Ankit Kumar (Scientist-C), with major improvements in functionality, scalability, and user experience.

This project used primary socio-economic and production data collected from a local farmer to assess crop profitability, cost structures, and resource utilization.

The analysis produced data-driven recommendations for crop selection and market strategy, with the potential to increase annual revenue by nearly 25%.

A movie-ticket booking web application built with Flask, Jinja2, Python, SQLite, HTML, CSS, APIs, and data-visualization components.

A short group research project on the importance of communication and listening skills.