Speaker Interview: Dr Elliot Banks, Chief Product Officer, BMLL talks with Quant Strats
Dr Elliot Banks, Chief Product Officer, BMLL
First published in Quant Strats, August 2023
An exclusive speaker interview with Dr Elliot Banks, Chief Product Officer, BMLL where he discusses the biggest challenges facing data scientists in 2023, the use of ChatGPT by quant funds and how funds can get new systematic strategies into production quickly.
What do you think are the biggest challenges facing data scientists/AI experts/quantitative practitioners in 2023 and beyond?
Data, data and more data. The challenge for everyone in the world of data science, analytics and quantitative trading has been and will continue to be data. Expensive quant models are not effective without high-quality, clean, consistent data that is easily accessible and usable. That is a theme we have seen over the past few years, and we are continuing to see with the rise of Large Language Models (LLMs) and their increasing use across every industry.
Dr Elliot Banks is set to speak at Quant Strats on the panel discussion Understanding risk and regime shifts – predicting and quantifying market volatility and creating resilient quant strategies.
This panel discussion will take place on the 24th of October at 9:20 am and he will be joined by Gareth Shepherd, Co-Head Voya Machine Intelligence, Voya Investment Management and James Delaney, Director of Government Affairs, AIMA. You can find out more information about it by downloading our agenda.
ChatGPT is everywhere and being used everywhere. How do you see quant funds using this new technology and what advice can you give people using it?
ChatGPT is a great example of the power of modern machine learning, but the biggest driver behind any good machine learning model (whether in natural language processing or predicting market prices) is good quality data that allows users to quickly derive insights. These large language models have millions of parameters and are trained on billions of data points. Before using these models and techniques, firms should focus on their data ingestion processes and ensure that their data is as clean as possible. This will allow them to train models that are accurate and usable in the highly complex world of capital markets.
What is your advice to funds hoping to get new systematic strategies into production quickly and more often?
To develop effective investment and trading strategies, quantitative or systematic hedge funds must carry out extensive research and run complex computation and continuous model calibrations. They require robust infrastructure to do this. Building such a data engineering infrastructure in-house means first putting necessary hardware and technical support resources in place, and can be a significant investment prior to generating any potential returns. The quality of market data and connectivity to reliable data feeds also plays a key role in strategy calibration, plus consideration must be paid to the high cost of building a team of quant researchers to curate, cleanse and manage the various data feeds needed to optimise signal detection and deliver the alpha generation process.
Quantitative and systematic hedge funds need to access critical data engineering infrastructure within a reasonable timeframe to shorten their time to market, and at a reasonable cost. They need to streamline the set up and ongoing management of data engineering processes, and be able to continue these tasks going forward without risk of disruption. Accessing a ready-to-use, quant-built data science platform can help hedge funds to access the necessary research infrastructure to support their fund’s specific operation and growth from the outset, without high upfront costs and lengthy timelines.
Time is also critical for hedge funds looking to get new systematic strategies into production. Developing an in-house data engineering infrastructure requires extensive planning, resource allocation and implementation time. However, the time-to-market issues associated with building infrastructure in-house can be solved by using an out-of-the-box Data Science as a Service capability that provides quants with a high-quality derived data source. By leveraging such a service, quants can gain easy access to fulldepth order book, including Level 3 data, without the need for in-house data engineering.