Rajesh Baidya

I’m a Computer Science & Engineering graduate and AI researcher focused on machine learning architectures, temporal forecasting, and deep CNN-based computer vision. I build structured, reliable AI pipelines—from data preprocessing to evaluation—so results are objective and reproducible.

In geospatial research, I developed LSTM-based time-series forecasting models for sea level rise prediction using tide gauge data from coastal Bangladesh stations. I also created a DEM-based flood risk mapping framework for coastal vulnerability assessment, including feature engineering, missing value handling, and visualization systems for spatial analysis up to 2050.

In deep learning for medical imaging, I conducted a comparative evaluation of 7 state-of-the-art CNNs for automated dermoscopic skin lesion classification. I engineered a deployment-aware pipeline using an augmented ISIC dataset and achieved peak 93.5% test accuracy and 0.98 macro-AUC with InceptionResNetV2.

I bring a research-first mindset, with a strong emphasis on pipeline bias control, model evaluation rigor, and clear technical communication. I also serve as an IEEE conference Technical Program Committee member, supporting peer-review evaluation, novelty assessment, and academic integrity.