Python

Tool-driven analytics agent using LLM orchestration, schema-based tools, and a semantic data dictionary.

A comprehensive repository of 100+ Jupyter Notebooks covering linear algebra topics, including matrix decompositions, Gaussian elimination, eigenvalues, SVD, and interactive Python/Julia examples for teaching and learning.

Check out our YouTube channel https://www.youtube.com/playlist?list=PLBDUlnmEqyXD_llq6wETUqRkJGRiu8wFU for tutorials and walkthroughs of these notebooks.

AI Radiology Copilot (ARC) is a CNN model app to recognize and label brain tumors in MRI scans.

This is just a portfolio project repository. I want to show recruiters and my hopefully next coworkers something I am working on. That's why I used an ECS as architecture. Right now, the base algorithm should work correctly. In the future, I plan to add demos, examples, wrapper and so on.

Pyxel is a general detector simulation framework (https://esa.gitlab.io/pyxel). An easy-to-use framework that can simulate a variety of imaging detector effects combined on images (e.g. radiation and optical effects, noises) made by CCD or CMOS-based detectors.

The Open Energy Tracker is an open data platform for monitoring and visualizing energy policy targets.

🚀 Project Description

This project aims to empower farmers through AI, IoT, and data-driven automation, transforming traditional farming into a smart, sustainable process. It provides end-to-end guidance — from crop selection to profit generation — using intelligent technologies.

Our system integrates AI-based crop recommendation models that analyze soil, weather, and market trends to suggest the most profitable crops. IoT sensors monitor soil moisture, humidity, and temperature in real time, while CNN-based image processing detects plant diseases and tracks growth stages.

An intelligent Gemini AI assistant generates automated farming schedules for irrigation, fertilization, and maintenance, ensuring optimal yield with minimal manual intervention. The backend, built with Node.js and Express, manages sensor data and communicates with a cloud server for advanced analytics.

The system also includes AI-based hardware sensors that perform on-field edge analysis and sync with the central server for instant decision-making. The React + Vite frontend offers an intuitive dashboard for real-time visualization, alerts, and AI insights.

With features like predictive yield analytics, pest control recommendations, and market trend forecasting, this platform provides farmers a complete technological ecosystem for smart, profitable, and data-driven agriculture. 🌾

A python tool box for GPUMD simulations

aGrUM is a C++ library designed for easily building applications using graphical models such as Bayesian networks, influence diagrams, decision trees, GAI networks or Markov decision processes.

Benchmarking open source hydrological model Shyft: https://gitlab.com/shyft-os/shyft. This repo contains helper scripts for simulation setup, results analysis and preparation of figures for publication: O.Silantyeva and S.Huang “Benchmarking Shyft hydrologic model performance of streamflow simulations in mainland Norway"

La spécialité NSI (Numérique et Sciences Informatiques) au Lycée Clemenceau à Nantes.

Des cours de première NSI, essentiellement structurés sur des notebooks Jupyter.

This project contains all training code relevant to the IOCCG, Copernicus 2025 Multi-sensor Ocean Colour (MSOC) course, coordinated by EUMETSAT in collaboration with NASA and ESA.

FELiCS - The linear flow solver.

Developed at the Institute for Fluid Dynamics and Technical Acoustics at TU Berlin, FELiCS is an open-source simulation code designed for researchers and engineers tackling complex flow problems. Its finite element discretisation of the linearised flow governing equations enables accurate modeling of intricate geometries and diverse physical scenarios — from combustion dynamics to aerodynamic instabilities and water turbine flows and many more.

Versatile. Open. Research-Driven

FELiCS breaks away from the limitations of traditional linear finite-difference solvers, offering a modular, general-purpose tool trusted in over 30 scientific publications.

A two-channel deconvolution method with Starlet regularization. The code also features a PSF reconstruction algorithm.

Python API and CLI for remote submission and monitoring of Python applications on HPC.

Documentation | A library to read and write AGS files using Pandas DataFrames

“Schnittstelle zur Wahldatenbereitstellung”: An API and a UI to access data about elections concerning the Verfasste Studierendenschaft of the Universität Heidelberg. Contributions welcome!