4th Annual Machine Learning in Quantitative Finance
Understand how machine learning can be implemented more effectively within business operations and assess best solutions to deal with the challenges posed by machine learning.
ENTER YOUR DETAILS FOR THE CONFERENCE AGENDA
What are you currently focusing on within real-time series modelling?
Our current focus is to predict / nowcast stock excess return for the remaining of the continuous market intraday.
For example, at 9:30am, we know everything that happened to the market and individual stocks up till that moment, aggregating all that data, we want to predict for the stock’s excess return for the remaining of the day. This prediction will be updated continuously based on intraday stock returns and news updates on the stock. New sentiment serves a real time update on stock’s fundamentals, it gives us the data points to continuously performing pricing discovery.
So our focus is short term nowcasting on excess stock return given real time stock data including news, returns, volatility.
In terms of technological investment, what areas are you focusing on? Where, for example, are you planning to invest within the next 3 to 6 months?
Such a great question. Quant trading and investing can’t be successful without a solid technology infrastructure that facilitates both research and executions. Over past four years, we built a solid real time streaming infrastructure that productionized our systematic market making, order matching and base level risk management automation.
Our next phase is to integrate alpha research and risk optimization into the execution. We want to build a pipeline that connects research to execution to continuously refine our alpha discovery and risk optimization, as well as enhance time to market delivery capability to be regime adaptative.
This requires to us keep expanding technologies in areas of real time streaming and big data technology that can handle both volume and velocity. We will be focused on building a quant trading research platform based on voluminous historical intraday data that can be calibrated with real time continuous data as well as integrating to our production execution environment in an agile way.
What are your main areas of focus at the moment regarding machine learning in quantitative trading?
As systematic equities market maker and risk manager, our business constantly needs to evaluate individual stock’s next move in the continuous market. We apply machine learning forecasting techniques to price moving real time data like news sentiment in addition to traditional market data to assess stock’s excess returns for short duration from hours to days.
What do you think are the current best practices for feature engineering in time series data?
We start with domain expertise – we only evaluate datasets that have highest short term pricing moving impact. The modelers in the space have in depth market knowledge as well as professional trading and investing experience. Domain expertise leads to a concentrated set of features. Given the time series nature of the data, we then add in reasonable delta, lags and aggregations to measure different signal durations as well as velocity and magnitude of changes.
Stock time series data doesn’t act in isolation – for example spread widening may have contemporaneous volatility increase – they are interactive. Being able to identify true lead / lag relationships versus random noise is an ever-evolving research subject.
We apply causal inference framework in time series study post selection based on domain knowledge, which provides expandability and stability.
Ahead of the GFMI 4th Annual Machine Learning in Quantitative Finance conference we spoke with Judith Gu, Managing Director, Head Equities Quantitative Strategist, US Equity Sales and Trading at Scotiabank. Judith Gu is the Head of Quantitative Equities Strategists overseeing North America equities quant trading for both US and Canada at Bank of Nova Scotia. Prior to joining Bank of Nova Scotia, Judith was a Quant Strat at Goldman Sachs working in both asset management and market making for 16 years. Judith holds an MBA in Finance / Technology from New York University, and an MS in Data Mining from University of Connecticut. Judith has also passed all 4 Certified Public Accounting exams.
October 3-5, 2022
New York, NY | Option to Attend Virtually
An interview with Judith Gu, Managing Director, Head Equities Quantitative Strategist, US Equity Sales and Trading at Scotiabank
Judith will be presenting during day two (10/4/2022) of the 4th Annual Machine Learning in Quantitative Finance conference!
For registration pricing and multiple attendee discounts, please contact:
Ria Kiayia
All Rights Reserved. marcus evans ® 2022
Interested? Do you feel you will benefit?
In terms of our platform, (our conferences are intimate peer-led meetings where all speakers and delegates are senior executives from top financial institutions), how do you see it assisting you with overcoming the challenges you currently face?
Learning and collaborating from outstanding peers from various institutions including buy side, market makers, and Fintech has been a highly invaluable experience for me. It let me keep up with up-to-date, enhanced research techniques. In addition, I am able to validate and collaborate ideas with peers as well as learn new ideas. On the technology front, I get to learn about best practices and more efficient technologies to solve challenging technological issues.
Case Study: Time series feature selection and modeling: Engineering, transformation and interaction
- Analyze how to apply domain expertise to feature engineering post basic data review rounds
- Discuss time series transformation: delta, aggregation, lagging – beyond auto correlation
- Time series interaction: where does causality come in?
- Modeling – does feature selection and model have to use the same technique?
Panel Discussion: Observations and evaluations around the current lack of ML/AI talent and education
- Why is there currently a talent gap in AI and ML?
- How is the current lack of talent directly affecting financial institutions?
- What are the best methods of addressing the current lack of talent?
- What impact is the lack of AI and ML talent having across the industry?
