Aram Salihi's Linkedin Analytics

I’m an AI/ML Engineer with a background in astrophysics (B.Sc) and computational science with a focus on biomedical computing and imaging (M.Sc, University of Oslo). My expertise spans machine learning, deep learning, image processing, and numerical modeling. Previously, have I worked as AI/ML consultant, but currently I work as an AI/ML engineer and researcher at the Norwegian Meteorological Institute. I develop AI-driven solutions for analyzing complex weather and climate data. More specifically engineering a graph-transformer based stretched-grid architecture (global and regional model) trained on reanalysis and benchmark against current NWP models. I also collaborate with ECMWF, pioneers in medium weather forecasting and also contribute to open source package called AnemoI which is an ai framework for creating data-driven climate and weather forecasting models. I also work extensively with high-performance computing (HPC/super-computer) and parallel GPU programming, using systems like (CSC) LUMI and (EUROHPC)Leonardo to scale scientific models and ML workflows. My main programming language is Python, complemented by experience in C, C++, and Bash. I have also a lot of experience with RCCL,NCCL, CUDA and ROCM, and cloud platforms such as AWS and Azure. I’m passionate about applying AI to real-world scientific and industrial challenges across disciplines, optimize models (speed,cost, memory,…) and deploy them in realistic hardware environments. GitHub: https://github.com/einrone Stretched-grid paper: https://arxiv.org/abs/2409.02891