Two top researchers at Jump Trading Group are leaving the quantitative trading firm.
Yiming Zhang and Darko Kirovski, who joined the proprietary trading firm in 2009 and 2011, respectively, resigned this week, the company confirmed.
Both were influential researchers in one of the largest and most profitable units at the firm, Jump Core Strategies, known for its prowess in high-frequency trading.
“Yiming and Darko were instrumental in Jump’s growth over the last 17 years, and we’ll be forever grateful for that,” Jump cofounder Paul Gurinas said in a statement to Business Insider. “They will be friends of the firm forever.”
It could not immediately be confirmed where Zhang and Kirovski are headed next. Zhang declined to comment, while Kirovski did not respond to a request to comment.
In a bio tied to a 2023 public speaking appearance at his alma mater, UCLA, Kirovski said that “he has had as much fun as any human in research has ever had at Jump Trading’s jcs group, one of the biggest market makers on the planet.”
“Interestingly, for more than a decade the group has been running probably the largest machine learning operation in the world unbeknownst to most,” the bio continues.
Chicago-based Jump is among a cadre of privately held, highly secretive high-speed trading firms that dominate many of the world’s largest financial markets. Similar to peers including Jane Street, Citadel Securities, and Hudson River Trading, Jump has grown substantially since the pandemic in 2020, when many of the market makers capitalized on the unprecedented volatility to print record profits.
The market volatility that pushed big banks to record trading revenues in the first quarter of 2026 has similarly benefited systematic market makers. Jump had its best quarter on record, one person familiar with the matter told Business Insider.
Founded in 1999, Jump now has more than 2,000 employees — up from 1,100 in 2022 — spread across eight countries and 13 offices, according to its website, and trades across most asset classes. Like its peers, it has also diversified beyond high-speed market making — high-volume trading at fractions of a second that add up to billions in annual profits — into strategies that involve holding securities for longer periods.
So-called mid-frequency trading, which has traditionally been the domain of quant hedge funds such as D.E. Shaw and Two Sigma, can play out over minutes, hours, or even days and relies on predictive signals derived from statistical research.
Many of the best quant trading firms have relied on machine learning techniques to develop trading strategies for years, long before the agentic AI craze ushered in by OpenAI, Anthropic, and other artificial intelligence labs.