Machine learning jobs are supposed to be the big new thing in financial services. After all, Goldman Sachs has created an elite new AI team, J.P. Morgan’s assigned ex-credit trader Samik Chandarana to develop machine learning strategies and has already unleashed LOXM, a new self-teaching trading algorithm, and UBS CEO Sergio Ermotti says 30% of banking jobs are due to dissolve because of this kind of automation in the next 10 years.
It might seem therefore that you should be positioning yourself now to chase machine learning jobs (even though Bank of America CTO Catherine says you’re already too late). But what if the finance jobs of the future are a lot less exciting than the finance jobs of the past?
This is a possibility raised by Saeed Amen in a new blog post. A systematic trader who’s worked for Lehman Brothers and Nomura, Amen is a now a systematic trading consultant and writer. He knows a lot about machine learning and he says its uses are more mundane than you think.
“The basic idea of many of the techniques which underpin machine learning is find relationships between variables,” says Amen. “The difficulty with finance is that relationships tend to be less stable (financial time series are not stationary), and often we don’t have sufficient data,” he adds. Outside of finance, Amen says machine learning is best used for image classification – identifying whether a car is a car. Inside finance, this can be harder: a car always looks the same, but financial time series are mutable.
Because finance data is so tricky, Amen suggests that one of the best uses for machine learning in a financial context is simply the cleaning and processing of data. – The trick is turning raw, partly irrelevant time series into something that can in turn be used to create new algorithms. His own company, for example, has created an index to measure the sentiment of communications by the Fed. The biggest part of constructing this index was, “collecting together all the Fed communications and speeches and sorting, before doing any sort of natural language processing or index construction.” For the moment, this kind of processing is still done ‘manually’, but ultimately Amen says machine learning algorithms should be able to discern the most important parts of text and ignore those that are not. This then, is the machine learning Holy Grail: data cleaning. Suddenly it doesn’t sound so alluring after all.
Separately, an equity researcher in the shadow of MiFID II has made an inexorably bleak utterance: “…this just feels like death,” he complained to Reuters in reference to his job as a researcher. He’d rather be working for a company in Silicon Valley: “It’s not even the money; it’s the optimism that I envy. Those guys are building a brighter future.”
Merrill Lynch thinks Barclays is going to have a good 2018 thanks to its investment bank, which has “a re-invigorated strategy, bigger balance sheet and new leadership team.” (Financial Times)
The attention and capital being lavished on Barclays’ investment bank have left some executives in the retail unit feeling sidelined. All the top managers in the markets business have been replaced in the past six months. (WSJ)
Michel Barnier pointed out that banks in London will lose the right to passport into the EU. (Bloomberg)
200 European Banking Authority staff are moving from London to Paris. (Independent)
“Our view is that the Government frankly is in chaos,” said a senior executive at a US bank. “We are really nervous.” (Independent)
Lloyd Blankfein on Brexit: ““Everyone needs a script, and fast.” (Guardian)
Six of the seven last U.S. Treasury heads have gone into finance. (Bloomberg)
Former trader who became a teacher: “The most difficult thing is the total loss of control. The timing of the day is completely set out, there’s no ability to decide to come in late…and during the day, every 55 minutes basically there’s a bell ringing.” (TES)
Rest before you are tired. Even if you love your job, (Medium)
Have a confidential story, tip, or comment you’d like to share? Contact: email@example.com
Bear with us if you leave a comment at the bottom of this article: all our comments are moderated by actual human beings. Sometimes these humans might be asleep, or away from their desks, so it may take a while for your comment to appear. Eventually it will – unless it’s offensive or libelous (in which case it won’t.)