How a quant billionaire is powering his trading models with ChatGPT-style artificial intelligence

Igor Tulchinsky made a fortune outsmarting human traders. Now he’s betting on large language models to make his hedge fund even smarter.

It’s a gray March day in Midtown Manhattan. Igor Tulchinsky, speaking from a wooden office chair between sips of coffee and long reflective pauses, lays out his latest algorithmic vision — integrating large language models (LLMs) into his fund, WorldQuant.
“The first thing an LLM can do is structure data — and 80% of all data out there is unstructured,” says Tulchinsky, dressed entirely in black, with piercing blue eyes shining with excitement. “It’s like a free lunch.”

Tulchinsky turned his long-standing obsession with data into a goldmine on Wall Street. The 58-year-old Belarusian émigré built a fortune in the hard, tech-heavy world of quantitative investing. WorldQuant, his fund, manages $10 billion for billionaire Izzy Englander’s larger hedge fund, Millennium Capital Management, from which he spun out in 2007. Another structure — WorldQuant Millennium Advisors, co-founded with Englander — manages an additional $13 billion for outside investors. Between his stakes in both companies and trading profits, Tulchinsky’s net worth is estimated at around $1.7 billion.

Quant traders like Tulchinsky write code that automatically executes trades based on pricing signals. WorldQuant’s specialty has long been statistical arbitrage — algorithms, or “alphas,” that exploit pricing inefficiencies between stocks or entire portfolios. These alphas engage in all kinds of trading — long positions, short selling, and more complex strategies. Whether it’s Trump’s trade policies or a tech company’s earnings report, WorldQuant likely has an alpha that profits from the situation.
“We trade the ripples, not the waves,” Tulchinsky says. “We have a warehouse with millions and millions of alphas.”

Now, he wants to supercharge this “alpha factory” by merging it with large language models (like ChatGPT), which he believes can generate new and better algorithms.
“We can use AI and LLMs to transform and discover alphas across different domains,” he says. “The possibilities are endless. LLMs are becoming more powerful.”
Though funds like WorldQuant have used AI for years — in research, forecasting, and coding — using LLMs specifically to create trading strategies is relatively new. In a 46-page Senate Homeland Security Committee report on hedge funds using AI (which includes WorldQuant, Citadel, Renaissance Technologies, and others), LLMs are mentioned only once.
“These proprietary LLM systems are an underrated factor in how investment firms might use AI,” says Francesco Fabozzi, director at the Yale International Center for Finance.

WorldQuant is well-positioned to capitalize on LLMs. According to Forbes sources, the firm employs over 150 PhDs in math, computer science, and STEM fields.
According to Fabozzi, a fund like WorldQuant might use an open-source model (like Facebook’s Llama) and train it with its internal algorithms to discover or even write new ones.
“You can take a standard LLM and enrich it with extra information. And we have a lot of internal data,” says Tulchinsky. “We can create a tool that answers extremely complex questions.”
Beyond alpha discovery, Tulchinsky wants to use LLMs to enhance research.
“You can ask it a question, give it a model like Ray Dalio’s, and when something happens — say Japan cuts interest rates — you can ask, ‘Okay, according to this model, what happens globally?’ And it’ll calculate it.”
This, he says, becomes internal data no one else can replicate.

Sources say WorldQuant’s use of LLMs is still in early stages but evolving rapidly.
To Tulchinsky, AI isn’t just a money-making tool — it’s a mindset, says retired four-star U.S. Army General Stanley McChrystal, who’s worked with him as an advisor since 2015.
“The ability to use data to anticipate events — that’s what drives him.”
Tulchinsky adds with a smile, “With the proliferation of data and AI, we’re reaching a point where you can quantify everything.”

His obsession with data likely stems from an unpredictable childhood. Born in Minsk in 1966, Tulchinsky grew up under the communist regime of Pyotr Masherov. His parents, both professional musicians, left the country in 1977 when he was 11.
“Back then, once you decided to leave, you were immediately fired and publicly labeled a traitor,” he recalls.

The family spent three months in Italy before being granted refugee status in the U.S. They moved frequently — across four states before he turned 17.
“It’s a transformative experience to give up everything you took for granted. You become less afraid of change.”
Fascinated by computers from a young age, he was programming video games by 17.
He earned bachelor’s and master’s degrees in computer science from the University of Texas, followed by an MBA in finance and entrepreneurship from the Wharton School. His first job was at AT&T Bell Labs, before transitioning to finance via Timber Hill — an options trading firm founded by another Eastern European émigré, Thomas Peterffy.
“Of course I remember Igor,” Peterffy says. “He’s a deep thinker. Sometimes he seemed completely lost in thought.”

In 1995, Tulchinsky began working with Izzy Englander, a young hedge fund trader who had founded Millennium. Tulchinsky earned a reputation as a statistical arbitrage expert.
In the early 2000s, they discussed starting a separate firm within Millennium.
“As a pioneer in quantitative investing, Igor has always been a step ahead,” says Englander in a rare statement.
But WorldQuant had a rocky start. In August 2007, quant funds crashed. WorldQuant suffered too, but Tulchinsky pulled capital before the bottom hit.
“Cutting losses is a core principle — that’s what risk management is,” he says. “People don’t do it because it’s emotional. But it’s a good strategy.”

As assets grew, so did the demand for global talent. The first international office was in China. Tulchinsky hired five out of a thousand applicants — “spectacular alpha creators.”
Today, WorldQuant has 1,000 employees in 27 cities across 16 countries, including London, Tokyo, Armenia, Hungary, and Vietnam.
“We enable talent, and it gives us alphas,” Tulchinsky says.

Outside work, between 2013 and 2023, he donated $65 million to his foundation, which supports WorldQuant University — an online school offering free master’s degrees in financial engineering.
He’s also invested in over 100 startups through WorldQuant Ventures. He’s authored three books on investing and is working on a fourth — focused on cutting losses.

Tulchinsky still has more to say, but the meeting ends. He has calls to take and an afternoon flight to Miami. He clearly doesn’t like staying still — the motto outside his office reads: “Change is progress.”
And with a parting comment, he says: “I looked at Belarus on Google Earth — nothing’s changed in 50 years.”
Then he walks out.