Raoul-Gabriel Urma could have been a trader. He could have been a top technologist at an investment bank. Or he could have worked for Google. He could have done a lot of things, but instead he’s chosen to proselytize about the need to know Python and data science techniques if you want your trading career to have a whiff of a future.
“A lot of what traders do can be replaced,” says Urma, echoing a recent comment made to him by a developer in an investment bank. “A button can be clicked by an algorithm that uses historical data to determine when to trade.”
Even as algorithms take over day-to-day trading, however, Urma says humans will still have a niche. And that niche will be in bespoke data analysis. “The value-added comes from custom work. This is what’s harder to automate and what can be done with Python,” Urma says. “If you want to get an edge today you need to create new strategies with Python. This is why all the traders at boutique trading companies are increasing Python proficiency.”
It could be argued that Urma is biased. He runs Cambridge Spark, a training provider that works with companies to teach staff Python and machine learning. Around 75 percent of his clients are banks, some of them the major players. Urma says one big bank is working with him to train all its business analysts in the switch from Excel to Python. Another is training rolling cohorts of its developers in advanced skills in data science and machine learning. “These are popular classes – we had 400 applications for 20 places in the first one.”
Mastering Python and machine learning to the necessary level isn’t easy. “You can take an online course for a day or two, but actually being able to embed those skills in your day-to-day work requires a longer commitment,” Urma says. Some banks recognize this; he cites one major U.S. bank that’s allowing a select few of its developers to devote one day a week to his course over a year.
Does this mean you should give up on learning Python if you’re the sort of trader who’s getting up at 5 AM, working 12-hour days and has no time for anything else? No, he adds: “You need to be in the loop with what the new generation are doing. You might not be a Python expert, but if someone talks about machine learning and Python you should be able to converse and to know what’s possible, at a minimum.”
Some senior traders don’t appreciate career advice from technologists. But Urma has the kind of impressive CV that should make even the most status-conscious person on a trading desk take note. He gained a first-class Masters in Engineering in computer science from Imperial College in 2011, followed by a PhD in computer science from Cambridge University, all before the age of 24. Along the way, he spent four months working for Google and six months at Goldman Sachs, but he ultimately decided that neither “big tech” nor a bank career was for him.
“I worked for Google in California. The salary and working conditions were good and you could have anything you wanted, but it didn’t match my goal in life,” says Urma. “I wanted to have a more direct impact, so I decided to start my own company.”
Meanwhile, banking was too bureaucratic. “I finished my Goldman Sachs six-month industrial placement when I was 19 and I figured it wasn’t the right path because things were too static for me,” says Urma, recalling that it took him nearly three weeks just to get a meeting where he could explain an idea for a new project to a manager. “In a start up, you have the idea and you can implement it by the end of the day!
“Big institutions don’t always have a culture of change. They lean more towards: ‘If it works, why change it?’”
This article originally appeared in eFinancialCareers.