Portfolio

This page lists a couple of things I’ve built or been a part of building.

HyPer is the main-memory database system prototype build at TU Munich. I did my PhD as part of the HyPer group during the first years of the prototypes existence. Lots of information as well as papers about the HyPer project can be found on the project website. In essence, HyPer is a modern database system which uses only main-memory to store data. HyPer compiles all transactions and OLAP queries to machine code using LLVM. Long-running queries can be executed concurrently with short transactions at incredible speed.

Together with a colleague, we submitted a solution to the challenge in 2013 which was ahead of all other submissions by about 30% for the final public evaluation. Unfortunately, the parameters in the non-public evaluation changed drastically yielding a second place out of about 60 participating teams for the hidden evaluation. Our code is public on GitHub.

During my PhD, I wrote multiple tools to facilitate teaching. On of them is Codematch which allows students to answer SQL exercises online and compete with each other. During the first year of usage, about 350 students used the tool for learn SQL. As I helped teaching the big database classes in Munich, I wrote a tool which helps managing exercise groups and exams the students take. I have published the code on GitHub together with a couple of screenshots.

I created a bunch of other tools to help teach students, for instance simple interfaces to run xquery, datalog and assign hundreds of students to exercise groups using linear programming (not public).

I have built many sideprojects over the years, some recent ones are:

Traffic Observer (Node) to determine whether I should live in San Francisco and commute or live somewhere closer to work.

docker-ssh (bash) to ssh into docker containers by name.

Also, feel free to take a look at my GitHub account for more.