Hoda Hashemi

About — CV

Open to research collaboration33.90°N — 35.48°E · Beirut

Hoda Hashemi

Computational Physicist · American University of Beirut

Hoda — /huː.dʌ/ — “who?-duh!”

My work sits where continuous physics meets discrete computation: translating the equations of geophysical flow into stable, fast, inspectable numerical systems — and building the tools that make those systems easier to drive.

Education

  1. 2023 — present

    M.Sc. Computational Science

    American University of Beirut

    Optimization, GPU computing, statistical learning, finite elements, vortex methods.

  2. 2019 — 2023

    B.Sc. Theoretical Physics · Minor in Pure Mathematics

    American University of Beirut

    Analysis, differential equations, algebra, manifolds, probability, numerical computing.

  3. 2021 — 2022

    Exchange Program

    Pitzer College, Claremont CA

    Statistical physics, French, and R programming.

Research

  1. 2025 — present

    Graduate Research Assistant

    Microflows & Microscale Heat Transfer Lab, AUB

    Numerical modeling of microscale flow and heat-transfer systems.

  2. 2024 — present

    MITgcm Simulations of Atmosphere–Ocean Dynamics

    Research project

    Simulation workflows and numerical experiments in climate dynamics.

  3. 2024 — present

    Vortex-in-Cell QGSW on the Sphere

    M.Sc. research

    Computational modeling of quasi-geostrophic shallow-water equations over spherical geometry.

Selected projects

  • 2025

    Agentic System for MITgcm

    LLM/RAG interface for querying documentation, modifying code, and executing simulations in natural language.

  • 2024

    Needleman–Wunsch on CUDA

    Parallel implementation on CUDA + Octopus HPC with reduction trees, stencils, coarsening, occupancy optimization.

  • 2023

    Diffusion–Convection PDE on a T-Membrane

    Finite-element and Galerkin analysis in FreeFem and MATLAB with numerical–analytical comparison.

  • 2023

    Inference of Nonlinear Dynamical Systems

    Data assimilation for the Van der Pol oscillator via adjoint-based BFGS and Bayesian optimization.

Stack

Languages

  • Python
  • Julia
  • Fortran
  • C/C++
  • MATLAB
  • R

HPC

  • CUDA
  • Slurm
  • MPI basics
  • Profiling

Scientific

  • MITgcm
  • FreeFem++
  • FEM
  • Vortex methods
  • Data assimilation

Tooling

  • LLM/RAG
  • Git
  • LaTeX
  • Linux