Piet Vinnbruck

Software Development - Data Science

Welcome! I am Piet Vinnbruck, a trained IT Business Administrator and postgraduate student in Information Systems. I offer customised services in the areas of software development and data science.

About me

My path into the IT world began with an training as a Management Assitant for IT Systems, where I gained valuable experience in consulting and process automation. I then decided to study Information Systems at the University of Münster. During my studies, I specialised in data science and software development and was able to put my knowledge to practical use in several projects.

 

Today, I support companies in optimising their business processes through innovative software solutions and data-driven analyses. My goal is to deliver tailored solutions that are not only of high technical quality but also make economic sense.

Services

service-01

Full Stack Software Development

I develop customised applications that can be seamlessly integrated into existing systems. I offer complete front-end and back-end development and create robust APIs for efficient communication between different services. From simple web applications to complex integration solutions, I ensure that your software is both functional and user-friendly.

Data Analysis and Predictive Modelling

I turn your data into valuable insights and support you in making informed decisions. This includes analysing and visualising existing data as well as developing predictive models using machine learning to identify trends at an early stage and optimise processes. My solutions combine data-driven analyses with modern machine learning techniques to develop future-proof strategies for your company.

Portfolio

GEMS GUI - Development of a User Interface

This project was developed as part of a project seminar at the University of Münster. The aim was to develop a graphical user interface for the GEMS software, which is used for modelling epidemics.

 

Technologies:

  • Angular CLI (Version 16.2.8)
  • HTML, CSS, TypeScript
  • MongoDB

Summary:

  • User-defined components for intuitive operation
  • Visualisations
  • Connection to a database

Face recognition with a Siamese CNN

This project was developed as part of the ‘Data Analytics 2’ course in the Information Systems Master’s programme. The aim was to develop a model for face recognition under the restriction of a maximum of 200,000 trainable parameters.


Technologies:

  • Python, PyTorch
  • Jupyter Notebooks
  • Optuna for Hyperparameter-Tuning

Summary:

  • Implementation of a Siamese Convolutional Neural Networks (CNN)
  • Comparing the performance of two Loss-Functions: Contrastive Loss und Triplet Loss
  • Use of synthetically generated pictures for training
  • Evaluating the model with the LFW-Cropped-Dataset

Contact me now

contact@pietvinnbruck.com

Do you have an idea for a project or are you looking for support in realising it? I look forward to hearing from you!