Case Study - Quantitative Research in Financial Trading

The purpose of this executive guide is to provide a brief overview of the typical data science methodology for financial services use cases of Machine-learning (ML)/ Artificial Intelligence (AI) within Capital Markets and Banking. For this use case, we specifically look at providing an overview and guide to an architectural solution for Quantitative Research within financial services – most commonly Capital Markets and Financial Trading.

In capital markets, quantitative researchers and electronic trading groups have developed algorithms using reinforcement learning to find the most profitable trading strategies and to optimise trade execution. Reinforcement could be based on simulation data and deep reinforcement learning (amongst various other techniques).

This paper will discuss how you can use AI to aid your quantitative research in Capital Markets Trading, along with other use cases such as Delinquency/PD, Spread determination, Cash flow prediction, Fraud, and regulatory compliance such as Trade surveillance for financial crimes.

Using AI software and bespoke hardware solutions, you can use AI to detect statistical anomalies and develop quantitative research techniques. Modern AI models are configured by human experts (as opposed to expert development) and then trained repetitively on a continuous feed of data. GPUs are key for training as well as inferencing (scoring) models on domain specific finance data.

Working closely with Nvidia as their Elite Solutions Provider, we have deployed thousands of GPUs for financial trading. We are proud to have been named Top Performing Tesla Partner in the UK for 2020 and Nvidia AI Champion for 2021.

We can provide a proof-of-concept (POC) service with VPN access to the Nvidia DGX A100 or special request for onsite delivery of an appliance, along with availability for AI technical sessions with a Deep Learning Engineer and discounts available at NVIDIA's Deep Learning Institute.