If you want to be assured of an enduring career in financial services today, it helps to be in with the teams bringing machine learning to banking. Banks are widely held to be laggards when it comes to the use of artificial intelligence, and are trying hard to catch up. Goldman Sachs alone, for example, is currently advertising around 100 jobs that specify machine learning skills globally.
Not all banks are equal when it comes to machine learning, however. And not all machine learning teams have been able to achieve results. - The field is fraught with internal politics as machine learning specialists from academic backgrounds find it difficult to implement their ideas in the face of resistance from established quant traders and engineering teams. "It's hard to say who's actually making money from this," admits one leading machine learning specialist at a U.S. bank. "There's a lot of spin."
Nonetheless, banks are trying. Oliver Blaydon, head of quant, risk and data science at headhunters Armstrong International in London, says the change has been perceptible. "At the latter end of last year, advanced analytics wasn't really a topic of conversation in banks. Now, growth is very clearly happening and some exciting tools are in development, although we are still a long way from using machine learning tools to full automate trading."
In the process, Blaydon says banks are coming to realize that they can just take a great quant and call him or her a data scientist: "You need proper machine learning expertise. Quants are great mathematicians, but they are coming up against the limits of their knowledge."
Based on conversations with headhunters and machine learning insiders, the people and teams listed below are at the forefront of developing the banks of the future. If you don't have a PhD in particle physics, you may at least want to learn their names.
We've listed J.P. Morgan's Intelligent Solutions team first because we're told it was one of the best in the industry. Formed around 2014, its stated intention at the time was to, "to transform and leverage JPM proprietary data assets into opportunities," and to offer, "strategic insights through data mining, analytics and modeling." The team was led by Len Laufer, the multimillionaire founder and CEO of Argus Advisory and Information Services.
Nowadays, however, J.P. Morgan Intelligent Solutions is no more. Insiders and headhunters tell us the team is "dead" and that its members have been dispersed to J.P.M's operating businesses where many are allegedly contemplating joining hedge funds. Whether the latter is true or not, there's certainly been a rush of exits from J.P. Morgan's machine learning corps. They include Geoffrey Zweig, the natural language processing expert who quit J.P.M after around a year for Facebook in February, Graham Giller, who went to Deutsche Bank in March, David Fellah, who went to fintech firm ITG in London in April, and Rajesh Krishnamachari,who left for Bank of America Merrill Lynch in the same month. Laufer himself has also left and there are (unconfirmed) rumours he's off to Cerberus to join former JPM COO Matt Zames.
Who, then, should you get to know if you want to work in machine learning at J.P. Morgan? You could try Samik Chandarana, the former credit trader and J.P. Morgan veteran, was made head of data science and analytics in the corporate and investment bank in October 2017. You could also try Manuela Vesolo, J.P. Morgan's brand new head of artificial intelligence research and head of the machine learning department at Carnegie Mellon University. Both Vesolo and Chandarana report into Sanoke Viswanathan, chief administrative officer of J.P. Morgan's corporate and investment bank.
Alternatively, many of J.P. Morgan's former machine learning stars are now in positions to hire elsewhere.
Barclays is behind when it comes to machine learning, and is also trying to make up for lost time. This month, the Briitish bank hired Adam Kelleher, the former principal data scientist at Buzzfeed as director and chief data scientist for research. Based in New York and reporting to to Jeff Meli, Barclays' co-head of research, Barclays says Kelleher will be building a "global team," with expertise in, "sourcing, normalizing, and utilizing alternative data sets."
Those in the know say that French bank BNP Paribas is one of the market leaders when it come to implementing AI in a markets context. BNP's artificial intelligence (AI) team is led by Joe Bonnaud, its London-based head of global markets quantitative research, data and AI labs. Bonnaud says he's hiring quants and computer scientists for positions in London, Paris, New-York, Hong-Kong, and Singapore.
Credit Suisse is also making up for lost time in machine learning. As of last year, it began building "CS Labs" in the San Francisco Bay Area. Run by Jacob Sisk,former head of the payments data science team at Capital One, CS Labs is tasked with transforming Credit Suisse with "audacious" thinking. Sisk says he's looking for, "self-driven polymathic hackers with AI/ML, data engineering,quantitative social science, design strategy, functional programming or DevOps skills."
If you don't want to join Sisk in San Fran, you could always try George Htin-Kyaw, Credit Suisse's London-based head of engineering and cloud machine learning. Kyaw describes his team as, "an internal start up and consultancy for cloud based machine learning," which includes, "an on-premises machine learning lab."
Failing that, there's also the possibility of joining Anthony Abenante's new global execution services team in New York. Abenante joined Credit Suisse from KCG last August and has been appointed head of a new execution services unit that will combine program trading and electronic trading. He said last week that part of the unit's purpose will be to solve questions like, "How do we incorporate machine learning to help our sales traders be better at their job?" It's not clear, however, with Abenante will be developing machine learning tools directly.
Deutsche Bank already has an established quant investment solutions team with a reputation for producing text analysis tools, but the man to watch there now is Graham Giller, who joined as head of a new primary research team in March 2018. A former head of data science at J.P. Morgan and Bloomberg, Giller is well-respected in the industry and can be expected to go looking for the best talent to join him. Giller is based in New York.
Goldman Sachs is widely held to be behind the curve when it comes to machine learning. However, it's hard trying to catch up. In November last year Goldman created a new research and development team, headed by Neema Raphael, a Goldman Sachs strats veteran who was previously responsible for building SecDB, the firm's risk and pricing database. Raphael is hiring for his team. However, the Goldman R&D operation is deemed a rarefied "fellowship" entity by some in the industry, who say more rough and ready data analytics tools are emerging from the firm's FAST (Franchise Analytics, Strategy and Technology) securities team, led by Samuel Krasnik in New York.
Morgan Stanley has been busy on the fringes of machine learning for a long while. The U.S. bank launched 'Alphawise,' a research unit focused on customized hedge fund research, in 2008. Alphawise employs around 30 data scientists to support its equity researchers with data-focused insights based on machine learning. Globally, Alphawise is run by Angus Lund, a Morgan Stanley veteran based in London.
Morgan Stanley also runs a group that goes by the name of "Morgan Stanley machine learning," which describes itself as the bank's machine learning "center of excellence." The team is headed by New York-based Ambika Sukla, an executive director who says he's focused on, "applying machine learning techniques to algorithmic trading, risk management, operations and compliance, and wealth and investment management." Some of Morgan Stanley's more interesting machine learning activities are centered on Montreal. The Canadian city has developed a reputation for its machine learning expertise and Morgan Stanley's Montreal Technology Center is emerging as the hub for its machine learning initiatives and is hiring.
UBS has historically focused on the use of machine learning in research. The Swiss bank runs an "Evidence Lab" team which was set up in 2014 by Juan-Luis Perez, UBS's global head of research who joined from Morgan Stanley one year previously. Perez promptly set about hiring former Morgan Stanley colleagues for the new venture, including Evidence Lab co-Head Richard Hockley, senior scientist Agrit Agrawal, and head of analytics Joe Cordeira,
More recently, however, UBS seems to have decided that the Evidence Lab alone isn't enough. Earlier this month, it appointed Christopher Purves, co-head of fixed income trading, as head of a new "Strategic Development Lab" comprised of 80 people focused on the use of artificial intelligence on the trading floor.
Have a confidential story, tip, or comment you’d like to share? Contact: firstname.lastname@example.org
Bear with us if you leave a comment at the bottom of this article: all our comments are moderated by 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.)