Abstract

Data Engineering has traditionally perhaps not been the sexiest of topics, but the recent explosion of data – and the need to make sense of it – has spurred many new technologies. Quantitative hedge funds like GSA crucially depend on good and timely data, as the input to our trading algorithms, and we spend a lot of time building reliable and fast pipelines to achieve this. Very often we can make use of existing open source solutions. In this talk, we’ll give an overview of the existing GSA technology and what we want to do next. Some of the topics covered are event-based pipelining, reproducible and bi-temporal data, efficient and flexible data querying and distributed transformations, leveraging technologies like Apache Kafka & RabbitMQ, Apache Arrow & Parquet & HDF5 , Kubernetes & Nomad, Docker, Kong, Arctic/Mongo & SQL . We’ll also show the view from a bit higher up, namely what other technology teams exist at GSA and how all these pieces interact to ultimately turn data into trades.

Speaker bio: Joris Peeters completed his PhD in Applied Physics from Ghent University, Belgium, in 2011. He subsequently spent three years in Nuremberg, Germany, designing algorithms and software for computational fluid dynamics simulations, with a focus on aeroplane wing icing and race car engines. After moving to London, he joined Winton Capital Management in 2014 and subsequently GSA Capital in 2017. His time in the wondrous world of finance is mostly spent building technology for trading systems, data pipelines, risk calculations and historical back-testing. Outside of work he enjoys quantitative sports betting and anything to do with computational neuroscience. He also has a passion – but sadly no talent – for playing the trumpet.