January 9, 2018
How Machine Learning Is Heating Up Tech Opportunities In Financial Trading

I was recently part of a panel at the Global Derivatives Conference here in Chicago in which the topic was the application of machine learning. Interestingly, it became clear during our discussion that there are several people who believe machine learning is, basically, the be-all-end-all “Holy Grail” for success in quantitative finance.

The real question is: Is it? Like most things, the answer isn’t quite that simple. However, what we can say is that we’re in a new phase in which machine learning is getting a supercharge. Machine learning encompasses a set of tools that are being developed at a dramatic pace. As part of this phase, people are trying to figure out how to apply new tools to the centralization of financial trading. To me, the interesting thing is that from everything I've seen, developing those applications is not necessarily going to be easy.

Any time you approach a new tool or technology that’s truly revolutionary, you’re going to have a philosophical discussion that asks, “How will applying this tool to a variety of things yield new knowledge for us?”


The Current Landscape As We Know It

Right now, you have large companies built on the central application of these technologies for financial trading. The most visible of those companies is called Sigma and a more prominent “stealth” company, Renaissance Technology, seems to have been applying all of these tools for a while, becoming one of the most successful of its kind. Beyond this tier of companies, more and more people are trying to get into this area. With almost any technology related to financial trading, there may be initial success followed by increased spending and ultimately, continued growth.

We’ve seen this pattern of initial success-spending-growth before many times, haven’t we? Take computers, for example. Over time, computers got to be so powerful and expensive that I've been told that at one point, some of our clients were replacing their PCs every three months because the PC coming out three months later was that much faster. That kind of speed truly made a difference to them.

So initial success leads to increased spending on technology, which eventually begs the question related to growth: How much do you want to spend to get another one dollar of net expected revenue? If we’re talking about spending one dollar, that’s an obvious decision. If it’s spending $1000, however? Not so obvious. If it’s $10,000, maybe it’s not worth the trouble after all.

As machine learning moves beyond the “early adopter” phase, we see quite a few companies seeking to continually invest in it that will probably shape the changes in the market. Surely this means we will see more companies succeed as others wither.

With increased investment in this area, what implications are there for the types of candidates that might be working within these companies on machine learning on the software side?

We’re beginning to see a lot of possibilities on this front. For example, we’re seeing the potential usage of image processing in analyzing satellite photos or photos taken by drones that, of all things, could be used in a financial trading capacity.

How’s that? Let’s say you’re a retail analyst and you want to know how much activity is happening on a particular day, like Black Friday, the day after Thanksgiving. In order to understand what the foot traffic might be to your store, you could analyze the number of cars that enter all the parking spots designated or near your store.

What if the parking lot is full? Well, then you might analyze if the cars parked in those spots right now are the same cars that were parked in those spots one hour ago. So to the degree that you can figure out how much traffic a particular store generated in terms of people coming in and out, you might obtain some significant insight to help you in planning going forward. You can then potentially say, “Our store did really well on this particular day, which happened to be when we had a blowout sale” or even “Our parking lot was busier this year by 10% compared to last year – what did we do differently this year?”

As another example, we know that as it enters the retail grocery space, Amazon has been working on a prototypical store called “Amazon Go” without any cashiers. The customer walks in, is identified by an Amazon Go app, selects the items they need and then leaves – without ever going to a register. How are they charged upon leaving the store? The system keeps track of the entire experience based on machine learning and a variety of sensors.

People are working on technologies just like this to provide greater financial insight so customers can get data that is ever-more quantifiable. However, here’s where a lot of opportunities may arise for candidates – with so many rollouts of new technologies, there is always a troubleshooting, testing, improving and evolving component that requires the human element. Machine learning is very exciting but it’s not perfect. What looks great on paper has to be executed flawlessly in the real world. As we know, that doesn’t always happen, thus the opportunity that exists for candidates.

What’s also only to the candidate’s benefit is that most forward-thinking firms in financial trading (and beyond) who will win in this highly competitive landscape will surely recognize that machine learning and its progression as it pertains to the company will require a lot of sustained investment over the long haul. They know it is a marathon, not a sprint. They also know that if they want to succeed and have a competitive advantage, the edge they want to achieve over other firms may be fully realized in the long-term rather than overnight.

If you find yourself interested in the opportunities that machine learning may present, pay attention to financial trading companies that are making such investments in these technologies. Yes, some are more secretive than others, but as more companies reveal themselves to be committed to achieving a long-term competitive advantage, it may be a good time to discover if their goals mesh with your own.

Roy Talman & Associates can supplement this process of discovery by shedding light on financial trading firms that are making significant headway on the technological front today and where they’re headed tomorrow. It’s what our over 30 years of building relationships in the financial trading space can provide you. Together, we can pave a more well-informed path for your career, even when the field and its associated technologies are in a state of constant change.