Sourav Rai

In my current role as a Data Scientist at SitInCloud, I've spearheaded advancements in cyber threat detection, focusing on ransomware and potential intrusions, leveraging autoencoders and a hybrid method combining static and dynamic code analysis. My internships at SitInCloud and ESILV involved implementing and improving ransomware detection algorithms using deep learning techniques. At IRIF (CNRS and University of Paris Cité), I researched pruning for deep neural networks, creating optimizers for models like ResNet and WideResNet on CIFAR and ImageNet datasets. Additionally, I worked on quantization for deep learning at NXP Semiconductors and Inria, enabling high-performance models on embedded chips. For CEPAM, I built a feature-based SVM model for classifying 3D model datasets of prehistoric shells. My MSc in Data Science and AI from the University of Cote d’Azur provided me with a strong foundation in advanced deep learning, Bayesian learning, and optimization model-based statistical learning.