Summary
Hossein Kashiani is a PhD candidate and research-driven ML engineer with nine years of experience specializing in computer vision, multimodal AI, and biometric security. His work spans practical defenses and generative attacks—ranging from frequency-debiased deepfake detection and robust multimodal LLM anomaly detection to state-of-the-art face morph generation/detection that ranked highly in NIST evaluations. He has a strong publication record with recent papers at CVPR, WACV, and KDD, and builds solutions that emphasize robustness to domain shift, post-processing artifacts, and distributional changes. Notably, he combines generative models (StyleGAN, diffusion) with ViT-based ensembles and adversarial training to both synthesize and defend against realistic biometric threats. Based in Clemson, he brings academic rigor and systems-minded engineering to bridge vision-language research and deployable, generalizable models.
9 years of coding experience
4 years of employment as a software developer
Doctor of Philosophy - PhD Computer Vision, Doctor of Philosophy - PhD Computer Vision at Clemson University
Master's degree Electrical Engineering - Digital Electronics, Master's degree Electrical Engineering - Digital Electronics at Iran University of Science and Technology
Persian, English