Hi there! I’m Sebastian, a researcher and software engineer. My main research goal is to make implementing systems software as easy as possible. I’m convinced that this is only possible when the programming model is as simple as possible but at the same time expressive enough to enable semantic-preserving compiler optimizations as a foundation for high-performance runtime systems. In my work, I apply formal concepts such as the lambda calculus to dataflow-based practical systems for (big) data management, microservices and soon even operating systems.
I graduated in 2008 with a Masters degree in Computational Engineering from the University of Technology, Dresden. My studies focused primarily on database, data management and distributed systems. During that time, I was a teaching assistant at the Chair for Database Systems under the supervision of Prof. Wolfgang Lehner who was also the supervisor of my Bachelors Thesis. My Masters Thesis centered around an online checkpoint/restart algorithm for dataflow-based data processing systems under the supervision of Michael Beckerle (back then IBM) and Prof. Christof Fetzer.
Already in 2006, I moved to Cambridge (US) to join IBM as a software engineer. After a year as an intern, I became a lead engineer for the design and implementation of a next generation data integration platform for hierchical data under the supervision of Amir Bar-Or. My focus was primarily on the associated compiler and parts of the runtime system.
In 2011, I returned to Dresden as a research assistant in the area of systems engineering diving into the topics of big data and secure systems. During this time, I was a tutor for various seminars on systems engineering and distributed systems. But ever since I started the work on an own dataflow engine for my Masters Thesis and the experience gathered at IBM, I was convinced that dataflow is the final solution to concurrent and parallel execution. So I continued to intensively work on my system that I termed
Ohua ever since. In 2012, I came across Clojure and finally realized in order to make dataflow a commodity it needs to vanish as a programming model and become solely an execution model. This is where I developed the idea of the programming model that is key to
In 2015, I finally met Prof. Jeronimo Castrillon and could convince him of my ideas. So I very happily joined the Chair for Compiler Construction to pursue my private project Ohua as the main topic of my Ph.D. I graduated in 2019 with the title “Dr.-Ing”. During that time I deepend my knowledge on programming languages especially in the domain of functional programming languages (think lambda calculus) and compilers. This not only shifted my focus from Java to Rust and even Haskell but also enabled me to design and implement a lambda-calculus-based compiler for
Ohua featuring transformations that provably preserve the semantics of programs.
After a short but interesting hiatus from university at Huawei, I’m looking forward to join the Composable Operating Systems group lead by Michael Roitzsch at the Barkhausen Institute in February 2020 to discover new domains where
Ohua can serve as a vehicle to speed up development and optimize execution.