An Interview with a Head of Data Science

What inspired you to pursue a career as a quant?

I've always had a passion for finance. It's always been what I've wanted to do. At the same time, I've always been very interested in technology. I've been programming and coding since I was small, but my degree was focused on finance. I never knew I could bridge my passion for finance with the technological side of things until I came across the world of quant finance by chance. I did some research on it and a light bulb went on. Quant finance was the perfect bridge between what I've always wanted to do. I think it's a great field. There's so much depth to it, and so many interesting problems. 

How did the CQF program impact your career trajectory? 

When I started the CQF, I was working in data science. I was looking for something that could help me bridge that gap into the financial world. I started doing the CQF part time while working. Even during my pure data science work, the techniques I picked up from the CQF helped me, especially some of the advanced mathematical concepts. Then an opportunity arose at Prescient, and it turned out that the skills from the CQF were very well aligned to what they were trying to do. So, the CQF definitely helped me to get where I am right now. I still have my books sitting on my desk in the office and refer back to them often. The CQF is really helpful in terms of progressing your career and helping you understand how everything fits together.
 

Can you describe a typical working day in your role?

We don't have a typical day, which I really enjoy. There’s a lot of research involved, so most of our days are spent looking into model development and opportunity identification. My team specifically focuses on what we can bring into our investment process that's new and adds a competitive edge. We have a team methodology of failing fast. So, we'll pick up something, maybe a new data set or a new modeling style, and we'll try that for a week. If it doesn't show value at the end of the week, we scrap it and move on to the next thing. If we start to find value, then we dive in deeper. 

Once we have these models, we then have to figure out how we get them into our investment process and ensure they are robust. That's where good coding practices come in because you need to be rigorous about every single thing you do. At the end of the day, your model can be good, but if it goes down in production when the markets are changing, then that's much worse. I would say we spend about 50% of our time on the development and research side and the other 50% on just getting the models out in a robust manner. Then on top of that, there's the testing frameworks as well - making sure that there is actually value in the model. So, there's quite a lot of things that we get involved with on a day-to-day basis. 
 

What's one of the most interesting or challenging projects you've worked on? 

At present we've been looking into how we can get more information into our system. The more information we have, the more accurate predictions we can make. We already have a lot of data because we are very systematic, so we already use about 500,000 different indicators throughout the entire investment floor and that's a lot of very traditional financial metrics. 

What we are trying to do now is bring in something that's a bit untraditional. We looked at the Federal Reserve’s Beige Books. They're quite detailed documents (about 50-60 pages long). We then parsed out a lot of different financial indicators. So, as soon as the Beige Book comes out, that runs through our automated data science pipelines and analyzes approximately 200 different economic series. These are things like labor statistics, pricing statistics - all sorts of things like that. That's the easy part. The challenging part is - now we have 200 different series, we need to show where they lead to investment ideas and where they fit into asset classes. This is not easy because there are a vast number of different things you need to compare. We had to create new algorithms to spot where this series is correlated into our current investment process, how we could aggregate them, and what relates to what to get something that we could test. Then we ran through our very rigorous tactical asset allocation and ensured that it shows up on all our asset classes. Just identifying those specific series was a very interesting problem. We ended up using all sorts of advanced time series clustering methods to solve that. 

What skills gained from the CQF do you find most valuable in your role? 

The biggest thing is that the CQF helped me step into the financial world after working in a different sector. Investment processes are complex and diverse: there’s equity, multi- assets, cash, income, etc. Without the CQF, I would never have been able to catch up with the information that I needed to know to be able to do my job. The CQF prepared me to step in and be comfortable with all these terms and fundamentals. It would have been very difficult to just pick that same knowledge up naturally on the job. The CQF really did prepare me to work in a quantitative investment firm and work with these tools on a day-to-day basis. Also, it was practical. So, you pick up the correct skills for industry-based work.
 

How are you using AI in your current role? 

Firstly, as a data science team, we're trying to implement more agentic AI into our investment process. If we have spreadsheets or text documents that do not follow consistent structures - we can now analyze that. It opens up a world of more generalized datasets that we could never have dreamed of using before because there would have been too much manual work. We're also using AI in categorization engines. We're using it to get sentiments about certain topics. AI allows us to do all of this automated data science and to quite a good degree of accuracy. 

Then on the second level, we've started adopting AI across our investment floor as a working tool. You can use it as a sparring partner when you're figuring out new methodologies, for example. However, it really shines once you have your models and want to deploy them. That phase used to take a long time because of all the programming. We've gone from about a two week deployment cycle to two or three days. Our productivity has skyrocketed as a result. Of course, you need to be careful when using it. Whenever we use AI in our processes, there's rigorous monitoring. If we spot something strange coming through, it gets flagged immediately and we deal with that on a case by case basis. That's very important because it can hallucinate. I definitely see it as a tool that will become more and more influential in the future as it opens up so many possibilities. 
 

What do you think will be the next big topic for AI and quant finance? 

Firstly, I see a lot of opportunities in the investment space. There are a lot of processes that are still very manual, so I see AI playing a big role there. It will speed up all of the repetitive work that needs to be done by those teams. At the same time, on the regulation side, it will allow us to deal with all of that in a timelier manner, instead of having to do all of these deep dives manually every single time. I think AI is something that you can't really ignore - there's a lot of opportunities as long as you control it. That's the biggest caveat of all this. Plus, on the regulation and data governance side there are, of course, some caveats coming with the use of AI. We're doing a lot of research into how that impacts us as a financial service provider as we must ensure that what we do with the data is safe, secure, and explainable at all times. 
 

What advice would you give to someone looking to enter this field now?

Start doing small projects because that will give you something to show prospective employers and build your understanding. So, if you're going through a course, take one of the ideas that you're learning about, pull together a little project about it, and see what you can do with it. The most important skill that you can learn is, when you encounter a problem, the thinking around “how do I solve that problem”. It's something that you need to practice and learn. That's something I appreciated about the CQF because you get projects that you have to solve by yourself. I've always done this - that's how I got my first internship. Of course, your qualifications are also valuable, but when I interview someone for my team, the first thing I do is look at their repositories - what they've done, what their code looks like etc. - before I even look at what qualifications they have. That's a really important point and I think it's a viable way to get into the industry.