AI-powered fund takes top prize in Aaron Selber Jr. Hedge Fund Course

A team of students with a plan to use AI to find insights hidden in corporate disclosures won first place and the top prize of $1,400 in this year’s Aaron Selber Jr. Course on Hedge Funds. The class final, in which students pitched their fund strategies to a panel of guest judges, took place on Dec. 5.

The winning team, Xtract Capital Partners, proposed a fund with a systematic equity long/short strategy utilizing proprietary large language models to analyze textual changes in corporate filings. The strategy is designed to exploit a fundamental market inefficiency: While investors immediately process quantitative data included in financial disclosures, such as earnings and revenue, they often overlook qualitative shifts in corporate communications that often signal future performance changes.

“We were looking at linguistic changes in firms’ corporate filings and found that these textual changes in 10-Ks, 10-Qs and proxy statements often foreshadow future performance, but markets are slow in pricing them in,” said Avik Hegde (BSM/MACCT ’26, SLA ’25), the team’s CEO. “Our strategy involved reading 10-Ks at scale using LLMs, running a similarity-in-sentiment analysis, and then trading based on that.”

The team’s back testing on the S&P 100 demonstrated a 17% annualized return over 10 years with a beta of -0.09, suggesting market-neutral characteristics.

“We ran it against the Fama-French Five-Factor Model, and it was totally uncorrelated,” Hegde added. “I think that really showed this is a behavioral inefficiency that markets haven’t priced in yet.”

“They were really strong presenters, and I loved that their strategy was based on a research paper they had read for Professor David Lesmond in the Darwin Fenner Student Managed Fund course,” said Mara Force, professor of practice in finance and instructor in the hedge fund course. “They integrated what they’d learned previously and took it to a new level, turning academic research into something actionable and investable. The judges were all very impressed.”

This year’s class developed funds utilizing a wide range of strategies, from leveraging geopolitical disruptions and AI stock momentum to exploiting mispricings in biotech and dislocations in crypto-asset markets. Helping the students along the way were industry mentors who met with their assigned team throughout the semester to offer advice and critique their pitches. This year's mentors included Ron Bienvenu, managing partner of Noetic Investment Partners; Jacques Soenens (BSM ’93), founder of Great Gable Investments; Joseph De Perio, managing partner at Metabilia; and David Thomas, CEO of Equitas Capital Partners.

“I am consistently impressed with the quality of the presentations, the in-depth research and the preparation that all the kids do,” said Scott Hayes, vice president with Fidelity Investments, who judged the competition along with Keith Katz, retired investment advisor with Morgan Stanley, and Corey Von Allmen, partner at 801 West Capital Management. “I mean they really knock it out of the park.”

Hayes cited a pitch from students in a previous year’s course that involved tracking and replicating the trades of members of Congress, a strategy that now forms the basis of at least two ETFs.

“I’ve done this for four years, and typically a handful of students will reach out to me after class,” he says. “I know at least two or three have started hedge funds, so I’ve seen the success that these kids achieve.”

The Aaron Selber Jr. Courses on Alternative Investments, which include the hedge fund course and a course on distressed debt, were established in 2015 to honor Aaron Selber Jr. (BBA ’50), a member of the Burkenroad family and lifelong supporter of Tulane who had a special interest in alternative investing. Watching this year’s final presentations via Zoom were two representatives of the Selber family: Robert Autenreith (E ’74, MBA ’78), Aaron’s son-in-law, and Dewey Corley (L ’70), his longtime friend and business partner.”