J.P. Morgan has made another big hire in its artificial intelligence program – and with the poaching of Apoorv Saxena from Google, demonstrated that it is prepared to compete with the top levels of Silicon Valley in order to do so. Saxena will be head of AI and machine learning services, joining Manuela Veloso, who was appointed head of AI research last year.
Looking at the slate of AI related products that JPM have launched in the last 12 months, it’s clear that the new head will be managing a significant global franchise rather than a blue-sky product development shop. In June, the bank unveiled a virtual assistant for clients of its treasury business, which aims to learn client behaviour and JPM them navigate the payment system. Their LOXM system has been live for over a year applying machine learning directly to trade execution in equities. And the investment bank has a whole section on its website for AI and machine learning white papers and product outlines.
But this is not a bank-specific trend unique to J.P. Morgan; JPM has actually been losing artificial intelligence superstars almost as fast as it has been hiring them, with members of the former “Intelligent Solutions Team” leaving for Facebook, Deutsche, BoA and various fintech firms and hedge funds over the last few years. The Saxena hire should be seen as part of a wider movement toward automation and machine intelligence as a hiring priority.
Given that, it’s perhaps a bit surprising that if you search by keyword, JPM isn’t currently advertising any jobs with an obvious AI angle to them. But this in itself might be informative. The hiring action is taking place at the elite level, where jobs are not necessarily advertised, because part of the reason that you hire an Apoorv Saxena or Manuela Veloso (and part of the reason why competitors wanted to poach senior staff away) is that they are part of a network and can bring in follow-on hires to build out their own teams.
This is the way that senior level hiring has always worked in banking, and at the top levels of computer science too. The people who get hired first are the leaders in their field, and they are expected to know who the most promising researchers are at the level immediately below. Either through having worked together before, or through mutual acquaintances and personal recommendations, the top hire, if they are of the right calibre, attracts colleagues. So if you want to be in with a chance at these jobs, it is best to be part of a social network that includes one of the “rock stars”. We haven’t yet seen a whole-team move for a group of artificial intelligence bankers, but we shouldn’t be surprised if it happens in a year or two.
Separately, in the world of venture capital, it’s notoriously difficult to find a way into one of those social networks. An interview with Semil Shah of Haystack Partners and Lightspeed Ventures gives the key tip of “come bearing gifts”. Not in the sense of baked goods or monogrammed golf balls, but if you bring four or five potential deals, then you will do better than if you show up simply bearing a resume.
For a VC deal to count as a “gift”, though, it has to be more than the hedge fund interview commonplace of “best ideas, long or short”. Shah suggests that an idea only counts if you can demonstrate a solid connection to the company founder and a relationship which gives some credibility to your recommendation, rather than just saying “here’s an interesting company” and waiting years to find out if you were right.
And the way you build these ideas is by “hitting your beat”. At Lightspeed, the emphasis is on partners having almost a “journalistic mindset”, constantly developing sources, talking to people and working with and meeting founders. But you can’t just be a “schmoozer” – although connections and personality will get you so far, “there has to be a there there” in terms of investment skill.
Perhaps most interestingly, Shah talks about his own moves to establish a VC presence in Seattle, and attempts by other firms to look for ideas outside the Bay Area. In his opinion, one of the most important constraints on the company founders he talks to is “will they be able to put together a loyal team in this town?”.
Meanwhile
Researchers (of all types, but economists are apparently particularly bad) tend to heavily overstate their confidence in the validity of their results. Charles Manski, a professor of econometrics, is on a mission to persuade people to take uncertainty seriously and to be open about their error bars. (Bloomberg)
The technologist who developed UBS’s robo-advisor is going to AQR, as head of “client technologies”. (Financial News)
Companies don’t like it when people get allocated shares in their IPOs then sell for a quick profit, but that doesn’t mean that client confidentiality doesn’t apply. Someone on UBS’ Asian equity capital markets team forgot this rule and gave away the identity of a seller to Zhong An, the insurance company which floated last year. Zhong An then contacted the investor, and in the resulting investigation, UBS ended up suspending its head of Asia ECM. (Wall Street Journal)
Why banks are going to have to build the importance of populist politics into their business plans. (The Banker)
The FT series on the problems of the audit profession continues, with stories of audit partners who got too close to their clients. (Financial Times)
Juergen Fitschen, former co-CEO of Deutsche Bank, says in an interview that if he had the chance to do things again, he would have accepted state aid in the crisis. (Handelsblatt)
The Value Investor’s Club is the web forum that you can only join if you’re good enough, but where everyone can speak as an equal, even though some are junior analysts, some are deli store clerks and some manage billion dollar portfolios. (Financial News)
Groupthink among the fund management industry means that diversification between asset managers doesn’t always help you stay out of crowded trades. (Bloomberg)
Showing a creditable indifference to the possibility of being regarded as a bit of a stereotype, the former CEO of Deutsche Bank Americas will now be entering the world of professional golf, as the new CEO of the PGA. (Golf World)
The Australian banks were the big survivors of the financial crisis, rising up the global market capitalisation league tables and knocking on the door of the investment banking big league. Now their domestic franchises are under threat from the Royal Commission, as this excellent long-read summarises. (Bloomberg Businessweek)
Jens Weidman is no longer front runner to succeed Mario Draghi at the ECB … (Bloomberg)
…while Mark Carney has apparently been asked to stay on at the Bank of England. (Evening Standard)
Image credit: batuhan toker, Getty