Research · Machine learning

You build the software and systems that let a trading firm's researchers train machine learning models faster and more reliably, and you help choose the right method for each problem. You join the Machine Learning team at Jane Street in London, where models trained on market data help drive trading.

Machine learning means training a statistical model on large amounts of data so it finds patterns itself, here patterns in market data. The team uses many kinds, from the neural network, layers of simple calculations loosely modelled on the brain, to the random forest and gradient-boosted trees, which combine many small decision trees into one prediction. The seat is a bridge between research and engineering. You understand the mathematics of a model well enough to advise researchers, and you write the code that makes their ideas run at scale.