After studying theoretical solid state and especially metal physics in Göttingen, I moved to the Max-Planck-Institute of Metal Research in Stuttgart for my PhD. My diploma and PhD work was mainly on Hydrogen in Metals, a field which is gaining a lot of interest again right now.

After my PhD I started at Bayer in Leverkusen in the central computational chemistry department. My first projects were in environmental distribution modelling and in "Self-learning Systems and Artifical Intelligence" (in 1987!!!). In the following years, I learned a lot on QSAR/QSPR methods, force-fields, and quantum chemistry, and aside I even achieved a quite reasonable understanding of molecules and their properties. I realized the importance of solvation and the deficits of the available methods in that regards.

Being pointed to dielctric continuum solvation models, I happened to have the quite successful idea of replacing the dielctric continuum by a scaled conductor, which was the birth of the COSMO (conductor-like screening) model, which nowadays is available in essentially every quantum chemical program in the one or other form. 

A few years later I realized the deficits of COSMO - and all dielectric continuum solvation models, and as a consequence developed the COSMO-RS (COSMO for realistic solvation) model, a combination of COSMO surface polarities with statistical thermodynamics of pairwise-interacting surface segments.

COSMO-RS widened the application range dramatically to mixture thermodynamics at variable temperatures, which opened the door for chemical engineering applications. 

In 1999 I started my own company COSMOlogic in order to provide the COSMO-RS method in user-friendly software. During the folowing 20 years my team and I succeeded in developing the world-leading COSMOtherm software suite for the robust prediction of a wide variety of fluid phase thermodynamics parameters. We happened to win several of the property prediction blind challenges as the Industrial Fluids Properties Prediction Challenge and the SAMPL challenge.

I have currently about 180 scientific publications, with about 35000 citations, resulting in a h-index of 59. (as of Google Scholar)