In brief, this group derives value from terabytes of non-stationary, messy and unstructured data using intricate technical and ML techniques. They are research driven, but with a tangible end goal of finding money-making opportunities in the economy.
A central quantitative research group inside a systematic trading business is hiring research scientists. Team of ~10, working across the whole firm rather than under one desk, building models to work out what is actually happening in the real world from data that was never collected to answer the question.
Concretely, the problems look like: inferring a causal effect when the confounders are latent and the only proxies you have are images and text. Making a valid inference from data you collected adaptively. Estimating something that has to keep working as the world drifts underneath it. Deciding which of a hundred candidate datasets is worth three months of your time.
What we are looking for
- A doctorate, or a technical degree with real research behind it, in statistics, econometrics, CS, ML, EE or a quantitative science
- Depth in some of: causal inference, time series, experimental design, uncertainty quantification, high-dimensional inference, representation learning on messy data
- Strong Python, or R, C++ or Julia, and research code other people can actually run.
- No finance background required or expected. Many of the team came from the physical sciences.