Tadzio
Geospatial Data Scientist with 8+ years of experience, passionate about leveraging remote sensing technology to address planetary challenges

About

I have a broad interest in the application of geospatial and remote sensing technologies to address environmental and societal challenges at scale. My background is in Remote Sensing & Oceanography, but since then I have branched out into data science, geospatial analysis, and software development. I’m particularly excited by the convergence of physics-based simulations (especially radiative transfer), remote sensing and machine learning. Take a look at my portfolio, and if you’d like to collaborate, please don’t hesitate to get in touch!

My Projects

Check out my latest work

I've collaborated on a variety of projects with multiple research institutes and clients. Most of my work revolves around using optical imagery to monitor and analyze our changing planet.

Advancing water quality monitoring from space

Advancing water quality monitoring from space

Optical satellite imagery is used to monitor water quality from space. On behalf of a client, I implemented a physics-based inversion model to derive indicators such as chlorophyll-a from ESA Sentinel-2 satellite. The analysis focused on lakes and rivers in Germany and the United States.

Python
Lmfit
Scikit-learn
Xarray
Pandas
Docker
Underwater hyperspectral imaging

Underwater hyperspectral imaging

In this article, we describe the use of underwater RGB and hyperspectral imaging to map the seafloor at centimeter-scale resolution. From the ingestion of raw data to insights into biodiversity, biomass and ecosystem health. This technology demonstrates the potential for precise, large-scale carbon assessment of aquatic ecosystems.

Python
Rasterio
Pytorch
Xarray
Scikit-learn
Dask
Kedro
Docker
AWS
Airflow
Physics-based inverse modelling of multi- and hyperspectral observations

Physics-based inverse modelling of multi- and hyperspectral observations

A scientific paper published during my PhD, presenting the development and validation of a physics-based model for both forward and inverse simulations. The model enables the simulation of optical signals measured by multi- and hyperspectral ocean-observing satellites.

Python
Lmfit
JAX
Scikit-learn
Xarray
Pandas
Hyperspectral image processing at scale

Hyperspectral image processing at scale

In this article, I outline the steps involved in processing large volumes of hyperspectral data, including the essential groundwork required before the data can be used for further analysis. The work focuses on implementing radiometric, geometric and water column correction models.

Python
Rasterio
Xarray
Scikit-learn
Dask
Kedro
Zarr
Docker
AWS
Airflow
Terraform

Collaborated with

University of Amsterdam
SCRIPPS
planblue
University of Queensland
Gybe Inc.
Global Change Institute
Vrije Universiteit Amsterdam
University of Amsterdam
SCRIPPS
planblue
University of Queensland
Gybe Inc.
Global Change Institute
Vrije Universiteit Amsterdam
University of Amsterdam
SCRIPPS
planblue
University of Queensland
Gybe Inc.
Global Change Institute
Vrije Universiteit Amsterdam
Contact

Get in Touch

Want to chat or work together? Send me a message or book a free call:

Work experience

  • Data Scientist & Geospatial Consultant

    Planblue GmbH

    πŸ‡©πŸ‡ͺ Bremen, Germany

    β€’ Implemented automated validation workflows and quality control for raw sensor data and end-user data products. β€’ Specified requirements and implemented solutions for geospatial data processing and storage using Xarray and Zarr. β€’ Delivered analysis-ready geospatial data products as GIS compatible raster files
  • Senior Data Scientist

    Planblue GmbH

    πŸ‡©πŸ‡ͺ Bremen, Germany

    β€’ Led the design and development of processing pipelines for hyperspectral imagery, transforming raw data into analysis-ready products β€’ Spearheaded the acquisition of a new optical sensor, enabling our imaging system to derive spectral reflectance measurements β€’ Designed and conducted calibration and validation experiments to ensure accuracy and reliability of hyperspectral measurements β€’ Interfacing between data science, engineering and business teams
  • Remote Sensing Consultant

    Gybe (Flying Gybe Inc.)

    πŸ‡ΊπŸ‡Έ Portland, Oregon, United States

    β€’ Implementation of a physics based retrieval framework (HYDROPT) for the retrieval of water-quality parameters. β€’ Implement the HYDROPT framework in a Docker container to scale and optimize processing of Sentinel-2 imagery in the cloud. β€’ Validation of water quality retrievals for lakes and rivers in the US and Germany against in-situ measurements.
  • PhD Remote Sensing & Oceanography

    Vrije Universiteit Amsterdam

    πŸ‡³πŸ‡± Amsterdam, The Netherlands

    β€’ Developed and validated a physics-based inversion framework for multi- and hyperspectral sensors (HYDROPT) β€’ Developed processing pipelines for hyperspectral radiative transfer simulations β€’ Published research and presented at international conferences.
  • Guest Researcher

    University of Amsterdam

    πŸ‡³πŸ‡± Amsterdam, The Netherlands

    β€’ PhD guest researcher conducting research on in-water radiative transfer and bio-optical models
  • Visiting Research Scientist

    Scripps Institution of Oceanography

    πŸ‡ΊπŸ‡Έ La Jolla Shores, California, United States

    β€’ On-going development of a physics based retrieval algorithm for satellite ocean color applications

Skills

python
xarray
numpy
scikit-learn
pytorch
rasterio
docker
geo-pandas
terraform
git
kedro
aws
airflow
plotly
dask
zarr
netcdf
c++
Training & Outreach

To learn is to grow. To teach is to give

Here you'll find a selection of workshops and training sessions I've attended, events where I've presented my work to wider audiences, and courses and workshops I've taught.

Training

E

3D Radiative Transfer Modelling workshop

Nov 2024
European Space Agency (ESA)
ESA ESTEC, Noordwijk, The Netherlands
Attended the 3D Radiative Transfer Modelling workshop as part of ESA's Spaceborne Imaging Spectroscopy event in 2024
O

Better metrics for algorithm assessment

Oct 2018
Ocean Optics XXIV
Dubrovnik, Croatia
More information can be found here
I

Sentinel for Water Resources

Sept 2017
Institute for Electromagnetic Sensing of the Environment
Sirmione, Italy
Topics included: Copernicus, ESA Sentinel 1, Sentinel 2 and Sentinel 3 missions with focus on water resources, water quality measurements and analysis, and operational applications. Furthermore: emergency mapping (flood event case study), water quality and fieldworks, ocean color radiometry, biogeochemical modeling, hydrological cycle and related parameters.
E

AMT for Sentinel Fiducial Reference Measurements

Jun 2017
European Space Agency (ESA) & Plymouth Marine Laboratory
Plymouth, United Kingdom
The AMT4SentinelFRM project focuses on providing high quality Fiducial Reference Measurements (FRM) to validate satellite data during the Atlantic Meridional Transect (AMT) annual research voyage between the UK and destinations in the South Atlantic.
T

Statistical Learning and Data Science

May 2017
Tinbergen Institute
Rotterdam, The Netherlands
Course on Statistical Learning and Data Science, taught by renowned statistician Trevor Hastie
Contact Me

Send me a message