Nitesh Kumar Shah

Nitesh Kumar Shah

M.Tech researcher in Data Science & AI at IIT Madras. I work on generative models, computer vision, and applied machine learning.

Publications

Journal Articles

Integrating FinBERT Sentiment with Financial Stress for Optimized Portfolio Performance

Published in Academy of Marketing Studies Journal, 2025

A portfolio-optimization approach that fuses FinBERT-derived market sentiment with financial-stress indicators to improve risk-adjusted returns.

Recommended citation: S. Pandey, R. Singh, N. K. Shah, and A. Bidhan. (2025). "Integrating FinBERT Sentiment with Financial Stress for Optimized Portfolio Performance." Academy of Marketing Studies Journal.
Download Paper

A Comprehensive Bluetooth Security Audit Framework for IoT Devices

Published in Computers & Security (Elsevier) — Q1, SCI [Under Review], 2025

A security-auditing framework for Bluetooth Low Energy in IoT devices, built on BT-HCI packet inspection and automated communication analysis. (Under review, Q1 SCI.)

Recommended citation: K. Leelasankar, N. Aslam, A. Vemuri, J. Gadde, U. S. Lavu, N. K. Shah, and S. Venkatesan. (2025). "A Comprehensive Bluetooth Security Audit Framework for IoT Devices." Submitted to Computers & Security (Elsevier), Q1, SCI. [Under Review].

Conference Papers

RainGAN-Kathmandu: A Generative Adversarial Framework for Synthetic Rainfall Augmentation in Urban Road Scene Datasets

Published in IEEE TENCON 2025, Malaysia, 2025

A GAN framework that synthesizes realistic rainfall over urban road scenes to augment datasets for robust perception in adverse weather.

Recommended citation: N. K. Shah, J. Gadde, N. Singh, C. P. Maurya, and S. K. Singh. (2025). "RainGAN-Kathmandu: A Generative Adversarial Framework for Synthetic Rainfall Augmentation in Urban Road Scene Datasets." IEEE TENCON 2025, Malaysia.

A Novel Adaptation of the UNIT Framework for Cross-Domain Climate Data Translation with Enhanced Feature Preservation

Published in 12th Int. Conf. on Signal Processing and Integrated Networks (SPIN 2025) — Springer, 2025

An adaptation of the UNIT (unsupervised image-to-image translation) framework for cross-domain climate data, with enhanced feature preservation.

Recommended citation: N. K. Shah et al. (2025). "A Novel Adaptation of the UNIT Framework for Cross-Domain Climate Data Translation with Enhanced Feature Preservation." 12th Int. Conf. on Signal Processing and Integrated Networks (SPIN 2025). Springer.
Download Paper

An Efficient ADAS Framework for Cross-Domain Image Translation and Feature Preservation Utilizing CycleGAN

Published in CONFLUENCE 2025, Noida — Springer LNEE, 2025

A CycleGAN-based ADAS framework for cross-domain image translation that preserves task-relevant features for driver-assistance perception.

Recommended citation: A. Bidhan, N. K. Shah, R. Kumar, V. Gupta, and S. Prakash. (2025). "An Efficient ADAS Framework for Cross-Domain Image Translation and Feature Preservation Utilizing CycleGAN." 15th Int. Conf. on Cloud Computing, Data Science & Engineering (CONFLUENCE 2025), Noida, India. Springer Lecture Notes in Electrical Engineering.
Download Paper

Datasets

Kathmandu Road Dataset: Ratnapark to Tripureshwor

Published in Harvard Dataverse, 2025

An open urban road-scene dataset covering the Ratnapark–Tripureshwor corridor in Kathmandu, published on Harvard Dataverse.

Recommended citation: N. K. Shah, J. Gadde, A. Bidhan, K. Gullapalli, C. P. Maurya, and N. Singh. (2025). "Kathmandu Road Dataset: Ratnapark to Tripureshwor." Harvard Dataverse. https://doi.org/10.7910/DVN/38E8AG
Download Paper