Shahab Uddin Email and Phone Number

• Independently managed multiple projects simultaneously.• Acquired experience in data collection, cleaning, and preprocessing, ensuring data quality and integrity for analysis.• Executed machine learning model training, evaluation, and validation, enhancing model performance and accuracy.Military Sealift Command Project: • Proposed an attention-based generative adversarial network for ship deck segmentation.• Innovated a generative approach for document enhancement using small unpaired data. Generated a large, paired dataset from a small unpaired dataset for training denoising models.• Demonstrated the effectiveness of synthetic datasets for training machine learning models for document denoising.Tissue Ablation Analysis: • Conducted tissue ablation analysis using CNN. Demonstrated that the bipolar nanosecond pulse is the optimal waveform for tissue ablation.Flood Detection Project (DARPA-funded): • Contributed to the post-processing of the flood detection results, improving F1 score and PR-AUC value by approximately 2%-13% for all methods.Advanced Research in Vehicle Detection and Image Translation (Funded by the US Army):• Attention-based Generative Adversarial Network Development: Developed an attention-based generative adversarial network for converting optical videos to infrared videos, achieving superior FID (2.20) and KID (32.26) scores, and enhancing detection percentage at 2000m by 6.17% using synthetic MWIR dataset.• MWIRGAN Development: Designed a generative model named “MWIRGAN” for visible-to-MWIR image translation, surpassing several state-of-the-art models.• Synthetic Image Generation: Suggested a domain adaptation-based method to generate synthetic MWIR images from game-engine-generated images.Machine Learning Model Security Analysis Project:• Conducted in-depth analysis of internal feature maps of deep learning models to identify adversarial model manipulations attacks on saliency map-based interpretation methods like Grad CAM, LRP.

• Assisted in teaching graduate and undergraduate courses in machine learning and discrete-time signal processing.• Prepared questions, graded papers and tests, ensuring fair and timely assessment of student performance.• Delivered presentations on various topics, explaining complex concepts in an accessible manner.• Supervised students' projects, providing guidance on research methodology, data analysis, and project presentation.• Supported professors in curriculum development and classroom management.

• Utilized physics-inspired machine learning to generate synthetic data, addressing data scarcity issues for radiation mapping.• Applied advanced machine learning techniques to enhance the accuracy and reliability of radiation mapping.• Collaborated with a multidisciplinary team.