Ari Granevich

- Led research on diffusion models for computer vision and signal processing, developing novel architectures for denoising and object detection in challenging domains - Designed and deployed deep learning models for image classification and live object detection, collaborating with Carnegie Mellon University researchers (published: "Using Radar for Edge-based Live Learning") - Built end-to-end ML pipelines using PyTorch, implementing data preprocessing, augmentation strategies, distributed training infrastructure, and large-scale evaluation frameworks - Directed team of 5 engineers on innovation project, driving generative AI research from concept to production deployment with measurable business impact

- Developed computational algorithms and aerodynamic models using Python and MATLAB for advanced systems analysis - Applied machine learning techniques to optimize performance and efficiency, building foundation in data-driven engineering - Gained expertise in algorithm development, numerical methods, and large-scale computational modeling

- Instructed and mentored 77 students in advanced flight physics, holding weekly tutorial sessions and office hours - Developed course materials, graded assignments and tests, and provided technical guidance on complex aerospace engineering concepts

(Direct Track from MSc) Researching the intersection of signal and image processing with generative AI. My work focuses on diffusion models, transformers, and large language models (LLMs). Advisors: Prof. Haim Permuter, Prof. Joseph Tabrikian, and Prof. Igal Bilik.

Thesis: Deep learning for heavy tailed signal processing. Supervisors: Joseph Tabrikian, Igal Bilik and Haim Permuter