If you’re a managing director at Goldman Sachs with a reasonable chance of making partner, why would you leave? Most people in finance would be intensely thrilled to get a sniff at a Goldman partnership. Unless, of course, something even better was on offer.
For David Ha, former co-head of Japanese rates trading at Goldman Sachs, and a managing director at 32, the better alternative was a residency at Google Brain, Google’s 12 month programme in California to jump-start careers in machine learning.
Ha declined to speak to us for this article, but his public profile shows him leaving Goldman Sachs after eight years in May 2016 and joining Google a month later. Colleagues say he quit. Ha was doing ok at Goldman: he was promoted to MD after six years with the firm and was then made co-head of Japanese rates trading a little while later. If he’d hung around, he might have made partner circa 2020.
Instead, Ha joins the ranks of ex-banking types building new careers in technology. In doing so, he could be strengthening his finance career for the future. One day, machine learning is expected to be integral to trading in hedge funds and banks. This week’s NIPS machine learning conference in Barcelona has attracted recruiters from the likes of Citadel and AHL.
Google Brain’s residency is notoriously difficult to get onto: thousands of top PhDs globally apply for around 25 places each year. So, how did a Goldman rates trader in Japan make the cut?
It helps that Ha has a Masters in Mathematics. It helps too that while he was at Goldman he ran a machine learning blog and was active on Github – where Google often approaches promising machine learning talent.
Does Goldman need to add Google Brain to the list of hedge funds and private equity funds trying to poach its front office hires? Not necessarily. Pay for the residency is thought to be around $100k, and although Google Brain is open to participants with ‘diverse backgrounds,’ insiders say there aren’t any other ex-Goldmanites there.
Long-term though, machine learning is likely to become more attractive to anyone in finance with a quantitative bent. “Finance is of miniscule importance in the grand scheme of things,” says one Brain resident. “There are far more interesting problems that can be solved with machine learning, from healthcare, to drug discovery, automated driving, robotics, and energy,
“The roles of doctors, lawyers, flow traders, middle management, accountants, economists, may be redefined or augmented with machine learning algorithms,” he adds. “A single professional worker in five years may have the productivity output of an entire team at present.”