Franz Srambical

Hi! I'm a researcher at p(doom) and research assistant at Helmholtz Munich (AIDOS lab), ex-Celonis software engineer, B. Sc. Informatics student at TUM and minoring in Computational Neuroscience (courses of M. Sc. Computational Neuroscience at LMU Graduate School of Systemic Neurosciences).

I am interested in creating systems that go beyond current language models and are able to reason and plan like biological intelligence. I am also interested in incorporating topological, geometric as well as physical constraints into learning algorithms. To be specific, I would love to create impactful learning algorithms that can be used throughout the drug-discovery process.

I would also love to play Mario Kart via my brain, which is why I am one of the founding members of neuroTUM. I am interested in representation learning for (read: including, but not limited to) EEG data. I believe that self-supervised pre-training of some form and subsequent fine-tuning will ultimately lead to useful motor-imagery classification algorithms.

I want to create AGI (i.e. design architectures that push the pareto-frontier of intelligent systems).

I have linkedin, twitter, google scholar and an unmaintained blog.

PS: no free lunch is a myth ;)