AI and Satellite Data for a More Sustainable Planet

I develop machine learning and deep learning models to extract meaningful insights from satellite and geospatial data for a more sustainable planet.

Nafiseh Kakhani

Where Vision Meets Impact

I am a Remote Sensing Engineer, currently an Earth Observation and Machine Learning Scientist at The Landbanking Group GmbH in Munich, developing the STOA model for remote sensing image processing in support of nature conservation and regenerative agriculture.

Previously, I was a Postdoctoral Researcher at the University of Tübingen under Prof. Dr. Thomas Scholten, applying computer vision, machine learning, and deep learning to Earth observation data for environmental insights, particularly related to soil. My background includes a Bachelor's in Geomatics Engineering, a Master's and PhD in Remote Sensing.

Remote Sensing

Extracting geospatial information from multi-source satellite data.

Computer Vision

Designing deep learning models for detection and classification.

Mathematical Modeling

Using optimization to solve complex environmental problems.

Selected Work

View all publications
XAI for SOC prediction
Explainable AI SOC Prediction SHAP Remote Sensing Machine Learning

Explainable AI for Soil Carbon

Interpreting soil organic carbon prediction models using a learning-based explanation method for transparent AI decisions.

European Journal of Soil Science, 2025

SSL-SoilNet framework
Transformer Self-Supervised Learning Soil Organic Carbon Deep Learning Sentinel-2

SSL-SoilNet

A hybrid Transformer-based framework with self-supervised learning for large-scale soil organic carbon prediction.

IEEE Trans. Geoscience & Remote Sensing, 2024

Conformal Prediction for SOC
Conformal Prediction Uncertainty Random Forest LUCAS Dataset Digital Soil Mapping

Uncertainty Quantification with Conformal Prediction

Quantifying prediction uncertainty of soil organic carbon estimation from remote sensing data using conformal prediction.

Remote Sensing, 2024

Python Python
PyTorch PyTorch
TensorFlow TensorFlow
GEE GEE
QGIS QGIS
CUDA CUDA
Docker Docker
Git Git
Linux Linux
Scikit-learn Scikit-learn
HuggingFace HuggingFace
NumPy NumPy
ArcGIS ArcGIS