I'm a full stack developer, passionate about building distributed systems that combine multiple technologies, components and approaches. My favorite work is to solve novel problems or to lower the barrier to solving existing problems.
Currently, I'm a Software Engineering Working Student at Microsoft, contributing to the Speech Services of the Azure platform. I work on the backend that orchestrates the model inference and fine-tuning for the publicly available speech-to-text products.
In addition to my professional work and studies, I usually also work on side projects to learn about new technologies and approaches. In this context, I'm primarily focusing on building solutions for problems that I can understand not only from a developer's perspective but also from a user's perspective.
March, 2024 - Present
Working Student Software Engineer - Microsoft
Mainly focused on developing the backend of the speech-to-text products that orchestrates and executes the model inference and fine tuning.
April, 2023 - March, 2024
Working Student Software Engineer - Avelios Medical GmbH
Contributed to the development of a modular hospital system. Mainly focused on developing web user interfaces and frontend features.
August, 2020 - April, 2023
Working Student Software Engineer - itestra GmbH
Contributed to multiple customer projects including the front and backend of a bank, the windows driver creation of industry sensors and the replacement of a legacy system for an insurance company.
Youtube Keyword Search
Commercial website that allowed to search youtube videos, playlist and channels word for word. Had ~10k monthly active users at its peak. Shut down after one and a half years of operation due to Youtube policy changes.
PeerGuard - Hackatum 2023, Winning Team
Disaster warning smartphone application that spreads emergency messages between smartphones via peer to peer connections in addition to the internet.
YChat
Instant messaging app with peer to peer video calling designed for reliability and scalability.
AR Robot Karol
Augmented Reality revival of the programming learning software "Robot Karol" for mobile phones
Telemedicine application
Telemedicine application suite for a sports medicine facility, allowing staff to monitor patent progress and activity between visits
Deep Learning based spine segmentation
Five layer UNet model that segments the vertebrae out of a 3D MR scan.
Implementation of a Boundary Attack on a Residual Neural Network
Implementation of a decision based adversarial attack to fool an image classification model to misclassify pictures.
Bachelor's Thesis - Secure Aggregation Framework for Mobile Applications
Framework for secure aggregation from mobile devices for federated learning
Master's Thesis - Opportunistic Networking on Contemporary End User Devices: Augmenting Emergency Notification Systems
Proposes, implements and evaluates the "Emercast" system that enables secure and efficient peer to peer emergency broadcast dissemination