Augustin studied Cognitive Computer Science and Intelligent Systems at Bielefeld University. Afterwards he spent two years building his own Start-Up with a Stipend from TU-Berlin and EXIST funding from HTW-Berlin.

He is interested in self supervised learning of structured representations that allow hierarchical abstraction and generative/predictive modeling at the same time. Similar to learned probabilistic grammars and hierachical clustering. He aims for approaches that can be framed as self-organizing systems, as this allows massive parrallelization and steers the learning rules towards biological plausibility.

Currently he is working on structured representations for double stranded RNA, to predict interactions with the ADAR enzym, which favours specific secondary RNA structure. Another interest is the extraction of protein sequence and structure motives for arbitrary downstream tasks, e.g. prediction of allosteric properties.

Personal fascinations are swarm behaviors, cognitive mechanisms in general and correlates of consciousness.