Daniel Shawul Abdi Email & Phone Number

As a scientific computing expert in engineering and atmospheric sciences, I have extensive experience in GPU computing, high-performance computing, deep learning, containerization, and cloud computing on AWS. I have worked on several projects that involved these technologies, including:At CIRES/NOAA1. Modernizing the Rapid Refresh Forecast System (RRFS)2. GPU acceleration the Common Community Physics Package (CCPP)3. OpenMP acceleration of the Community Radiative Transfer Model (CRTM) and GPU port of a compute-intensive kernel4. Docker/Singularity containerization of the Global Forecast System (GFS)5. Migrating GFS to the cloud6. Use of machine learning to predict soil moisture from GOES Advanced Baseline Imager (ABI) dataAt Tempoquest Inc1. Led the development of a GPU-accelerated WRF model2. Speedup of 5 to 7 times within a year of developmentAt Naval Postgraduate School1. Non-hydro-static Unified Model of the Atmosphere (NUMA)2. Acceleration of NUMA's algorithms on GPUs and Xeon-Phis3. Scalability of NUMA using 16384 GPUs of Titan supercomputer4. Unified Continuous/Discontinuous Galerkin implementationAt Western University1. Computational Fluid Dynamics (CFD) program from scratch2. High performance Computing (HPC) with MPI , Pthreads, CUDA/OpenCL, Cilk, OpenMP3. Structural Analysis and Design Program with FEM (static/dynamic analysis)4. Optimizing programs with Intel Vtune, Valgrind, Assembly level5. Scripting with Python, bash, tcsh, perl etcAt Florida International University1. Numerical wind mapping in complex terrain and built environmentPurpose: Micro-siting of wind turbines, Effect of wind speedup over hills and escarpments2. Model scale testing in boundary layer wind tunnels (BLWT) at UWOPurpose: Design of buildings for static/dynamic wind loads3. Full scale testing at Wall of Wind (WoW) for hurricane simulations at FIUPurpose: Mitigation of damages due to hurricane and other extreme wind conditions.4. Use of Artificial Neural Networks(ANN) for BLWT profile developmentAs a hobby, I enjoy applying deep learning for games, such as computer Chess and Go. I have witnessed all the major breakthroughs in AI for games, including the Monte-carlo Tree Search (MCTS) revolution for computer Go and the deep-learning revolution from Deepmind. I have also developed a chess program using ResNET for evaluation that was at some point in the top-5 in the world.With my extensive experience and expertise in scientific computing, I am always eager to take on new challenges and help organizations achieve their goals.

Listed skills include Fortran, Mathematical Modeling, Mathematica, Simulations, and 10 others.