This page provides a non-exhaustive list of my professional, academic, and personal projects. Almost all my professional projects are closed-source. Most of my academic coursework projects can’t be shared due to honor code regulations. An exception are projects that were suggested individually, and of course projects related to research.
Some of the projects mentioned below are publicly available on my GitHub repository.
- S-RASTER: Contraction Clustering for Evolving Data Streams: conceived the idea, implemented a proof-of-concept, mathematically formalized the algorithm and described it in a paper (Fraunhofer-Chalmers Research Centre for Industrial Mathematics, Oct 2019 to Mar 2020)
- OODIDA: On-board/Off-board Distributed Data Analytics: Developed a prototype of a platform for distributed real-time data analytics for connected vehicles, which was successfully deployed and executed on prototype vehicles (Fraunhofer-Chalmers Research Centre for Industrial Mathematics, Jan 2018 to Jan 2020)
- distributed data analytics: performed a comparative study of stream processing frameworks, created a prototype of a large-scale distributed system in an Internet of Things setting with tens of thousands of edge nodes, created a prototype of a federated-learning-based distributed machine learning framework (Fraunhofer-Chalmers Research Centre for Industrial Mathematics, spring 2017 to winter 2017)
- large-scale processing of telemetry data: implemented the data processing pipeline for large-scale processing, discovered RASTER, a very fast big data clustering algorithm, developed an application that produces a time-evolving network graph of clusters based on batch-processing of GPS data (Fraunhofer Chalmers Research Centre for Industrial Mathematics, winter 2016/spring 2017)
- energy consumption prediction in the smart grid: created a Markov process-based simulation of supermarket customer behavior, performed predictive statistical analysis of customer numbers and energy consumption (Chalmers University of Technology, autumn 2016)
- automating large-scale PLC programming: designed and implemented PLC Factory (source code), an application that removes a lot of the tedium associated with programming PLCs by dynamically generating files for TIA Portal (Siemens) and EPICS database records; a key component was a custom embedded domain-specific language, which I created (European Spallation Source, summer 2016); a paper is forthcoming
- machine learning: using a neural network and conformal prediction in order to identify regions with a potentially increased arrhythmia risk (Fraunhofer-Chalmers Research Centre for Industrial Mathematics, autumn 2015 to spring 2016)
- designing and implementing a framework for parsers of partly unstructured log files that meets soft real-time goals and exploits the power of manycore CPUs; (Jeppesen, summer 2015)
- designing and prototyping a static analyzer for configuration files, in the context of a greenfield project at Ericsson (Ericsson, via Dfind IT and HiQ, spring 2015)
This is a selection of projects I completed as part of my (formal) education in software engineering and computer science. I also completed a large number of projects in related online courses (MOOCs), but adding those would make this list rather unwieldy.
- Master’s thesis project (spring/summer 2016): evaluation of center versus edge stream processing based on the Linear Road benchmark (Java, Apache Storm)
- Independent Project in CS (spring 2016): Cryptographic block ciphers in functional programming languages.
- Parallel Functional Programming (spring 2016): fault-tolerant distributed MapReduce, parallelizing a Sudoku Solver (both in Erlang).
- Bachelor’s thesis project (spring 2015): Compiling Agda to Fω. Type-theoretic specification of the semantics for compiling Agda to System Fω.
- Advanced Functional Programming (spring 2015): Turtle graphics DSL in Haskell; research project: embedded DSL for a fast Fourier transform working on synchronous data flows (Haskell, Feldspar); collaboration with Niklas Logren. This extension was used for Markus Aronsson’s synchronous dataflow extension to Feldspar.
- Concurrent Programmming (autumn 2014): concurrent railway simulation (Java), distributed IRC chat server (Erlang)
- Programming for Mobile Platforms (summer 2014): wrote ‘WordFinder’, mobile application that is able to look up matching words based on a pattern, e.g. ‘f__b_r’ matches with ‘foobar’, but maybe some other words as well? The app incorporates a third-party API for looking up word definitions, and uses a persistent database. [Android, Java, SQLite]
- Programming Language Technology (spring 2014): parser for a large subset of C++, compiler for a C-like language (Haskell, JVM), interpreter for a functional language (Haskell)
- Project: Industrial IT and Embedded Systems (spring 2014): building an autonomous miniature vehicle, colloquially referred to as ‘self-driving car’, using OpenCV and OpenDaVINCI; group project, contributed by prototyping and implementing the main algorithms [Python, C++]
- Functional Programming (autumn 2013): Blackjack engine, Sudoku solver, Game of the Amazons (all in Haskell), collaboration with Benjamin Liebe. The implementation of the last project is available available on Github.
- Project: Systems Programming (spring 2012): several demos of which a pathfinding routine was arguably the most impressive one, targeting the educational robot Nao; group project, made a significant contribution regarding the design and implementation of algorithms [Python]
- Project: Programming (autumn 2012): decision support system, built in Java (GUI and program logic) and MySQL; group project; delivered basically the entire backend
In my spare time, I enjoy tackling problems with a functional programming language, preferably Haskell. One such project, an implementation of the game 2048 in about 90 lines of Haskell, got some attention on Github and Twitter. The full source code is available on Github.
- Several other Haskell projects, including some that are larger than my implementation of 2048, are accessible in my Github repository “haskell_playground”. Check out Evil Hangman!
- Several smaller programs are described on various blog posts. To access those, please click on the header “index”, and look for the section “implementations”.
- Even though the individual tasks were not particularly time consuming, the bulk of my solutions to the entire set of CodingBat exercises quite certainly constituted a larger undertaking. When I published my solutions in both Python and Java, they were the only publicly available complete set of solutions. I decided to publish mine because most solutions I came across online were fairly awkward. Here is the starting page to the Java solutions, and here the starting page to the Python solutions.