Chung (a pseudonym) is leaving Goldman Sachs’ strats team. After several years with the firm, he’s had enough. He’s going to try his luck with a hedge fund, or a fintech. Anything but a bank.
Chung was a sales strat (the Goldman Sachs term for quants) at Goldman: his role was to structure deals and analyze trades. If a client wanted to put a hedging strategy together using options, that was him. If a client wanted an analysis of the correlation between volume and relative value, that was him too. “I was basically a quantitative salesperson,” he said. “I used my quant skills to drive revenues, but it wasn’t easy. It only ever felt marginally useful to the franchise or clients and it didn’t feel like a sustainable career. The more senior you get, the harder it becomes to prove your value.”
Chung’s withdrawal from the Goldman strats team is quiet compared to the most voluble exit in recent history. That of Antonio Garcia-Martinez, a former pricing strat on the credit derivatives desk. Garcia-Marquez left Goldman in 2008 and wrote a vehemently anti-banking blog about the whole experience two years later. There, he argued that Goldman’s quants were a group of failed scientists working alongside “complete tools” in sales and “bat wielding gorillas” in trading. “We were basically the trader’s little bitches,” claimed Chung, adding that the embattled quants tried to maintain their sense of cerebral superiority by writing, “academic papers on the more theoretical aspects of their work,” – although their names were erased whenever they left to do something else.
The quants we spoke to said Garcia-Martinez’ experience applies to a lost past. Banks today are a lot less raucous: there are none of the food eating competitions he complains of. But there’s still a shortage of the rarefied academic pursuits that make quants feel special. Andrej Karpathy, a Stanford PhD and research scientist at OpenAI has just calculated the key institutions whose research papers have been accepted by ICML, a top machine learning conference coming up in Australia. The top twenty include Google, Microsoft and Facebook, as well as leading universities like Berkeley, Stanford and Princeton. There are no banks on Karpathy’s list. There’s not even a hedge fund. If the list is a proxy for institutions engaged in original research into artificial intelligence (which is what Karpathy suggests), finance looks pretty dire. This might by why David Ha, a former co-head of Japanese rates trading at Goldman, quit for a sought-after residency at Google Brain when he wanted to learn about machine learning.
Some quants aren’t even working at the algorithmic coal face. As a sales strat at GS, Chung says part of the problem was that he wasn’t actually writing code: “I might’ve stayed longer if I was.”
Garcia-Martinez didn’t respond to a request to comment for this article, but when he left Goldman he spent long hours coding. First, he created AdChemy, a bid management tool for online media exchanges. Next, he coded AdGrok, a search marketing tool. The latter was sold to Twitter and Garcia-Martinez went on to work for Facebook. Since 2015, he’s been writing a book about Silicon Valley; things have worked out very well.
For Chung, the future is less assured. He’s hopeful about hedge funds, although they’re more interested in data specialists that in quants who’ve worked in sales jobs. Maybe this is a bad time to leave Goldman anyway? Marty Chavez, the new Goldman CFO, is a former strat himself and the bank’s own jobs site is filled with strats roles as the firm pursues Chavez’ vision for a “data lake” overlaid by machine learning. If he were to stay at GS, which he won’t, Chung says he’d probably angle for a role coding the whole derivatives process. “That workflow is receiving a lot of attention – the entire, ‘client calls sales, asks for quote, sales calls trading desk, trading desk prices it, tells sales the offer, sales relays to client,’ thing is just begging to be automated.”