As a research assistant at the University of Greifswald, I conducted research in the field of object recognition on images using deep learning. During my research, I concentrated on improving Convolutional Neural Network-based image classification by exploiting information from the network layers. To improve the classification process, I particularly focused on detecting out-of-distribution data using this information. I used the following technologies: Python, PyTorch, fastai, Numpy, Pandas… Show more As a research assistant at the University of Greifswald, I conducted research in the field of object recognition on images using deep learning. During my research, I concentrated on improving Convolutional Neural Network-based image classification by exploiting information from the network layers. To improve the classification process, I particularly focused on detecting out-of-distribution data using this information. I used the following technologies: Python, PyTorch, fastai, Numpy, Pandas, Scikit-Learn, OpenCV, Jupyter, Git, Linux, k-Means, PCA, UMAP, Convolutional Neural Networks. In addition to research, I supported the following lectures with tutorials: "Introduction to Computer Science", "Software Engineering", "Computer Graphics" (using C++ and OpenSceneGraph), and "Evolutionary Algorithms" (using Java and ECJ). Show less

As part of the information extraction team at Gini GmbH, I worked on a system that extracts information from business documents (e.g., invoices) through named-entity-recognition and document classification using techniques from the field of linguistics and machine learning. Our team contained between 5 and 7 members. To organize our agile software development process, we used the Kanban method. My tasks included researching, implementing and evaluating various approaches to information… Show more As part of the information extraction team at Gini GmbH, I worked on a system that extracts information from business documents (e.g., invoices) through named-entity-recognition and document classification using techniques from the field of linguistics and machine learning. Our team contained between 5 and 7 members. To organize our agile software development process, we used the Kanban method. My tasks included researching, implementing and evaluating various approaches to information extraction and classification. I used the following technologies: Scala, Git, Linux, Gradle, Regular Expressions, Machine Learning, Test-Driven Development. Show less

As an intern in IBM's DB2 database development department, I supported the DB2 text search team and the DB2 infrastructure team. My responsibilities included writing tests, fixing bugs, evaluating text search features, and managing software builds for AIX and Linux.

During my studies at the University of Rostock, I worked as a student assistant in the computer science department. My responsibilities included checking the homework of the computer science students from the two classes "Abstract Datatypes" and "Algorithms and Programming" as well as supporting two research projects. The first project was a student project on vehicle routing for geospatial data using linear programming and evolutionary algorithms, while the second project was a university… Show more During my studies at the University of Rostock, I worked as a student assistant in the computer science department. My responsibilities included checking the homework of the computer science students from the two classes "Abstract Datatypes" and "Algorithms and Programming" as well as supporting two research projects. The first project was a student project on vehicle routing for geospatial data using linear programming and evolutionary algorithms, while the second project was a university project on decomposing linear opimization models for the distribution of gas supply. Both projects were implemented in Java and GLPK (GNU Linear Programming Kit). Show less