Physics-trained computational science student working at the intersection of scientific computing, Earth system modelling, climate data, and machine learning.

I am particularly interested in using numerical simulation and data-driven methods to study physical and environmental systems.

Current focus: Scientific Python · Numerical Simulation · Earth System Data · Machine Learning
Based in: Cologne, Germany
Open to: HiWi / Research Assistant · Internship · Scientific Computing · Climate & Energy · Scientific ML

Simulation & Modelling 2 · University of Cologne · Jul 2026

Numerically simulated the propagation of surface gravity waves using the two-dimensional shallow-water equations and investigated how discretization and grid design affect numerical stability and accuracy.

• Derived the simplified shallow-water equations under hydrostatic balance
• Implemented finite-difference discretization using an FTCS-based numerical scheme
• Analysed numerical stability using the Courant-Friedrichs-Lewy (CFL) condition
• Investigated grid-scale checkerboard oscillations produced on a regular grid
• Applied a Shapiro filter to suppress numerical noise
• Implemented an Arakawa-C staggered grid, placing surface elevation and velocity components at different grid locations
• Compared regular-grid and staggered-grid simulations under identical initial conditions
• Demonstrated improved stability and reduced numerical oscillations with the Arakawa-C grid
• Simulated the propagation of a localized surface perturbation with periodic boundary conditions

Tech & Methods: Python NumPy Matplotlib PDEs Finite Differences FTCS CFL Stability Shapiro Filter Arakawa-C Grid Shallow-Water Equations

Simulation & Modelling 1 · University of Cologne · Jan 2026

Developed and evaluated data-driven approaches for estimating precipitation from three-dimensional radar reflectivity data.

• Processed 3D radar volume data matched with rain-gauge observations
• Worked with strongly imbalanced precipitation data containing a large proportion of non-rain samples
• Compared three modelling approaches: an empirical Z-R relationship, Linear Regression, and a Graph Neural Network (GNN)
• Used graph-based modelling to represent spatial relationships between radar bins across elevation sweeps
• Evaluated predictions using RMSE, MAE, P95 absolute error, and bias
• Achieved improved predictive performance with the GNN compared with the empirical baseline
• Presented the final results to the supervising professors

Tech & Methods: Python PyTorch Geometric scikit-learn NumPy Matplotlib Graph Neural Networks Radar Data Regression Model Evaluation

Earth System Data Processing · University of Cologne · Feb 2026

Designed and implemented a reproducible Python workflow for processing multidimensional atmospheric datasets from raw input to analysis-ready output.

• Designed an end-to-end data-processing task from dataset selection to final output
• Accessed and processed atmospheric data in GRIB and NetCDF formats
• Used xarray and cfgrib for labelled multidimensional data processing
• Applied spatial subsetting, temporal aggregation, and regridding
• Converted processed datasets to Zarr for chunked, analysis-ready storage
• Structured the project as a modular and reproducible workflow using configuration files and documentation
• Used Git throughout development and submitted the completed project through a Git-based workflow

Tech & Methods: Python xarray cfgrib ERA5 GRIB NetCDF Zarr Git Regridding Reproducible Workflows

Programming & Scientific Computing
Python · NumPy · SciPy · Pandas · Matplotlib

Earth System & Geospatial Data
xarray · ERA5 · NetCDF · GRIB · Zarr · cfgrib

Machine Learning
scikit-learn · PyTorch · PyTorch Geometric · Regression · Graph Neural Networks

Numerical Modelling
ODEs · PDEs · Finite Differences · Numerical Stability · CFL Analysis · Numerical Simulation

Research Workflow
Git · GitHub · Jupyter Notebook · VS Code · Linux/CLI · Bash

Developing practical experience with machine-learning methods for environmental and geospatial data, including supervised learning, deep learning, feature engineering, and spatial model evaluation.

scikit-learn · PyTorch · xarray · Remote Sensing · Earth System Data

Deepening my understanding of numerical algorithms for physical systems, with a focus on discretization, stability, convergence, and translating mathematical models into reliable scientific software.

Numerical Methods · ODE/PDE Solvers · Scientific Python · Simulation

University of Cologne · Oct 2025 - Present

Relevant areas:
Earth System Data Processing · Simulation & Modelling · Machine Learning for Earth System Sciences · Physical Climatology · Numerical Methods · Statistics & Data Analysis

Izmir University of Economics · Graduated 2025 · Full Merit Scholarship Graduated 2025 · Full Merit Scholarship

Forschungsfabrik Mikroelektronik Deutschland (FMD) · Mar 2025

Five-day intensive programme focused on sustainable and energy-efficient ICT technologies.

Energy-Efficient AI · Distributed Embedded Systems · Life Cycle Assessment · Eco-Design

Scientific Machine Learning · Climate & Earth System Modelling · Numerical Simulation · Environmental Data Science · Battery & Energy Systems Modelling · Scientific Software

LinkedIn · Email