I am a Data Scientist / Engineer, currently working as consultant for Enterprise Data and Analytics at Capgemini. In my current role, my work is dedicated to analyzing client data, bringing alive data-driven software, implementing data pipelines and machine learning lifecycles, and designing cloud architectures. I focus on Natural Language Processing use cases and help clients to bring their AI prototypes to production by applying automation and MLOps tools.
Below, you can find a digest of the projects I am currently working on or have finished in the past.
| Name | Use Case | Tech Stack |
|---|---|---|
| Language Models | Experimental use case development and deployment based on Generative AI | Huggingface |
| Speech Transcription | Speech transcription using Azure cognitive services | Azure SDK, Streamlit |
| Named Entity Recognition | Recognize entities in text using the BERT model | Transformers, PyTorch |
| Name | Project | Year | Tech Stack | Description |
|---|---|---|---|---|
| Patent Classifier | Multi-Label Patent Classification with Deep Neural Networks | 2021 | TensorFlow, Transformers |
A comprehensive study to identify, implement and evaluate suitable approaches for the classification of patents using different neural network architectures like CNN, RNN, and Transformers. A domain-specific data set of 200.000 patent documents is used. |
| Crowdedness Prediction | Crowdedness Prediction in Public Transport Under Covid-19 | 2020 | PyTorch |
During Covid-19 pandemic social distancing in public transport is an important matter to prevent spreading the virus. Thus, it would be beneficial to know when and where there are bottlenecks in the public transport network. Our goal is to reduce the capacity problem by predicting the crowdedness for a specified time interval with RNNs. |
| Name | Description | Hosted Instance |
|---|---|---|
| My Portfolio | A dashboard to keep track of a stock portfolio's development based on the order history. | 📈Streamlit App |