After having completed my Bachelor studies in Software & Information Engineering with a focus on data mining I was still fascinated how to solve non-trivial real-world problems with computers using statistical algorithms.

The curriculum was very broad and covered aspects such as text mining, image understanding and signal processing to analyze acoustical data. In project-oriented group work we could get an in-depth view of many aspects in computational intelligence and machine learning. The flexible curriculum allowed me to strengthen my skills in statistics, functional programming and data visualization.

During my studies I developed an interest in analyzing financial markets using sophisticated machine learning (ML) models which was seen as ridiculous by most professors at that time. Considering the the low signal-to-noise ratio in financial data they were obviously right for some problems (without adding any model bias/domain knowledge) but I had the feeling that they just didn't want to cover any financial domain at all. I think that statistics is a very important foundation of computational intelligence and should get more weight in the curriculum.

Having worked as a quant at hedge funds and banks I was typically the only one with a CI background. However, skills like model design (e.g. avoiding overfitting), programming and creative thinking which I further developed during the CI studies clearly helped me in my job. However, I think that domain knowledge (like financial markets in my case) is absolutely essential and must be obtained on top of the CI studies. My first job was at Superfund Research where I developed automated trading systems on futures markets. Currently I work at a small proprietary trading firm in Zug, Switzerland.