I have worked on diverse research projects spanning multiple disciplines, ranging from Statistical Physics to Epidemiology to Computational Neuroscience.
Publications
Dynamics of platforms
I’ve analyzed how social media platforms work, and what can be changed in terms of algorithms and design decisions to improve them beyond engagement metrics. I approach this through large-scale data analysis and modelling.
Telegram is an important platform, given its size and use for less visible, more private content-sharing. We parsed and made available a large-scale multimodal (text, audio & video transcriptions) dataset of German-language telegram channels. The underlying 25TB multimodal ETL pipeline (Python, Snakemake, PyTorch, Whisper) achieved a ~250× real-time transcription factor per GPU on whisper/large-v3-turbo through finetuned batching and VAD.
Dynamics within platforms
I’ve analyzed how certain important societal issues interact with, and are affected by, social media. These include
how conspiracy theories evolve on German Telegram
how the discourse around climate change has evolved on Reddit
Computational Neuroscience
In my PhD, I worked on developing and validating physics-inspired models for large-scale brain activity. We find that coarse-sampled signals of neuronal activity are prone to spurious correlations and autocorrelated measures of neuronal activity should preferably use spiking signals.
Epidemiology
During the COVID-19 pandemic, we worked on analysing the effect of lockdowns on the initial spreading of the virus. This work was highly publicized, and impactful in the context of the pandemic in Germany.
Complex Systems
From my background in Physics, I’m interested in modelling of complex systems and networks in general. In particular, I’m interested in the interplay between dynamics and topology in evolving, complex networks.
2026
The Schwurbelarchiv: a German-language dataset derived from multimodal Telegram content for the study of conspiracy theories