Yiming Zhang and Darko Kirovski, influential researchers behind Jump Trading's massive machine learning operation, are departing the firm at the peak of its profitability.
Jump Trading just logged the most profitable quarter in its 27-year history, and two of the architects behind that success are walking out the door. Yiming Zhang and Darko Kirovski, senior researchers in the firm's elite Jump Core Strategies unit, resigned this week. Company cofounder Paul Gurinas confirmed the departures, calling both researchers instrumental to the firm's growth and noting they will remain friends of the firm.
The timing is jarring. First-quarter market turbulence driven by shifting trade policies and interest rate uncertainty pushed major banks to record trading revenues. Systematic market makers like Jump reaped the same windfall. Yet losing two veterans who helped build the engine is a meaningful development for a firm that thrives on intellectual property wrapped in secrecy.
Zhang joined Jump in 2009 and Kirovski arrived two years later. Both spent well over a decade inside Jump Core Strategies, the unit responsible for high-frequency trading across global markets. In a 2023 speaking engagement at UCLA, Kirovski offered a rare glimpse into the group, noting that for more than a decade it had been running what he called probably the largest machine learning operation in the world, largely unbeknownst to the public. That detail matters more now than when he said it. While the broader technology sector was still debating the commercial viability of large language models, quant firms like Jump had already spent years deploying those techniques at scale in live markets.
The exits come at a moment when Jump's competitive landscape is shifting. Jane Street and Citadel Securities have expanded aggressively in Asia, with Jane Street reportedly offering compensation packages north of $500,000 to incoming Hong Kong hires in late 2024, according to industry data referenced by Bloomberg. Jump itself has grown headcount from 1,100 employees in 2022 to more than 2,000 today across 13 offices and eight countries. But scale brings its own challenges, particularly in retaining the specialized researchers who give these firms their edge.
There is also the matter of where Jump has retreated. The firm paid a $123 million penalty in December 2024 to settle federal investigations into its crypto trading practices, followed by a civil lawsuit from the Terraform Labs liquidator a year later. The collapse of Jump's crypto arm, once among its most lucrative divisions, placed even more weight on Jump Core Strategies to deliver. Losing two senior minds from that unit raises legitimate questions about succession and whether the research culture they built can persist without them.
Neither Zhang nor Kirovski disclosed their next moves. In the current market, that silence speaks volumes. Top quant talent increasingly has the option to launch proprietary funds rather than remain inside large organizations. Researchers at Point72 and Two Sigma have made similar exits over the past year, departing established platforms to pursue independent ventures backed by institutional capital. The economics are compelling: why trade someone else's capital for a salary when you can raise your own fund and keep a share of the returns?
Jump's diversification beyond pure high-frequency trading into mid-frequency strategies, those holding positions over minutes or days, suggests the firm anticipated this kind of talent risk. Strategies that rely on statistical signals and longer time horizons require a different research skill set than ultra-low-latency execution. Whether Jump has enough depth on its bench to sustain both tracks without Zhang and Kirovski remains an open question.
For competitors and recruits alike, the signal is clear. The quant industry's talent wars are intensifying at exactly the moment when market volatility is rewarding the firms best equipped to exploit it. Watch where Zhang and Kirovski land next. Their destination will tell you as much about the future direction of quantitative finance as any earnings report could.
