Founded in 2018, Judge Research is the commercial extension of Dr. Nicholas Adams Judge's political economic research into applied problems in tech & finance.

Since 2021, that has taken the form of building AI models for systematic & market making (MM) strategies. After 6+ years of being nose-down in PyTorch, a class of MM world models have emerged as very profitable & scalable. Click live to see the latest iteration's track record - but only if you are feeling lucky; it is often down for training or other fun stuff.

The completion of this new class of models leads naturally to a MM-powered frontier research lab focused on world models and RSI.

Why Market Making as the Engine of a Frontier Lab?

(1) Financial markets are data-rich enough to support worlds models but enjoy learning challenges that are different from those of the video and physical world data which are the foci of most world model research labs. (2) MM strategy research should also of course be kept out of public training sets.

These two challenges necessitate a unique path towards RSI that can generalize to DL model-building in the hard sciences. Happily, because model improvements lead directly to higher profits, the {research methods ↔ pnl ↔ institutional capacity} feedback loop is very tight. As an aside, it would actively benefit from any potential market downturn.

Get in touch! Nick [at] judgeresearch [dot] co