You are now being redirected to Orchard Platform Markets, LLC.
If you aren’t redirected in a few seconds click here.
You are now being redirected to Orchard Indexes.
If you aren’t redirected in a few seconds click here.

Subscribe to our Newsletter


  • Investor Portal

  • Originator Database

  • Manager Database

If you forgot your login email, please contact your Orchard Account Manager or

Industry Profile – Hugh Edmundson

CEO & Co-Founder, Theorem 

“I’m not sure how familiar you are with machine learning,” speculates Hugh Edmundson — CEO and Co-Founder of Theorem — an investment management firm that uses data science and machine learning to invest in marketplace lending loans.  Edmundson proceeds to not only expound on machine learning, but how it’s driven a lot of the way he thinks about the problems he’s attacked in his career and the work he’s doing now at Theorem.

As an undergrad at Carnegie Mellon, Edmundson studied computational finance — a discipline that uses the tools of mathematics, statistics, and computing to solve complex problems in finance. While completing his degree, he also spent a substantial amount of time in the university’s world-renowned machine learning department. Machine learning, a subset of artificial intelligence, is the science of getting computers to perform an action without explicitly programming them to do so — essentially, teaching computers to think.

Edmundson began investing in marketplace lending while still an undergrad.  “When I first signed up for Lending Club, they were running a promotion where they would match whatever amount you deposited,” he says.  Having been a student at the time, he jokes that his first reaction was “free beer money” but quickly realized that the money was only accessible if used to invest in a loan — which he proceeded to do. The first loan he funded on the platform was for an individual who worked in IT at an investment bank and was looking to consolidate credit card debt. Seemingly a risk-free investment on paper, Edmundson posted the money and within several months, the borrower had defaulted. Finding himself net down on his commitment to Lending Club, Edmundson, who had taken his share of probability classes, knew there had to be a better way to assess risk systematically and programmatically to help identity quality borrowers.

Meanwhile, Edmundson graduated from Carnegie Mellon and began his career with Morgan Stanley, where he traded credit derivatives “as part of the post-financial crisis clean up crew.”  Basically, a bespoke synthetic CDO (Collateralized Debt Obligation) is a highly customizable security similar to a traditional CDO, but which utilizes credit default swaps or similar derivatives as its underlying collateral, rather than real assets. While at Morgan Stanley, Edmundson spent a lot of his time learning how to quantitatively assess risk in the corporate space and developing credit spread option correlation hedging algorithms using matching learning systems.  Although he was enjoying working for the bank, both from an intellectual and monetary perspective, he observed the marketplace lending space taking off and understood its huge potential.  “I saw parallels between making markets for corporate debt and what platforms such as Prosper and Lending Club were doing with consumer debt, but they were doing so more efficiently with APIs — I couldn’t stop thinking about it,” he says.

Edmundson arrived at the conclusion that the best way to get involved, given his background and domain expertise, was to start his own investment management firm dedicated to marketplace lending — in 2013, Theorem was born. He enlisted Carnegie Mellon classmate and “phenomenally smart guy” Abeer Agrawal as his CTO and co-founder. Agrawal had previously worked for Google before moving over to Mobilespan, which was serendipitously acquired by Dropbox around the same time Edmundson left Morgan Stanley.  What came next was an unconventional but critical step, which served as a major catalyst for Theorem.

Edmundson decided to apply to Y Combinator — the storied startup incubator that has helped fast-track companies such as Airbnb and Stripe. “I wanted to figure out whether I was completely crazy for thinking about this,” says Edmundson. Companies like Theorem are not the prototypical candidates for programs such as Y Combinator. “People usually associate Y Combinator with the venture cycle, but really it’s a community of founders and a network of advisors that help you think about common problems that any business might have when starting out. Besides, what are financial firms if not just technology companies that build technology to deal with dollars rather than searches or likes,” he says.

Theorem emerged from the three-month program with a strong foundation from which to launch its new fund. Not only did they secure investors, which included Ron Conway and Max Levchin, but they also developed improved strategy and operations. Edmundson says that going through Y Combinator has given Theorem a tremendous advantage in terms of culture and retaining the best quantitative talent. It’s also enabled San Francisco-based Theorem to recruit best in class and highly sought after software engineers, who Edmundson says, “typically would not give the time of day to a traditional investment firm.”

Theorem, which has commitments in excess of $150MM, is structured as a single fund and currently invests across consumer loans from quality originators that meet certain criteria, which right now include Prosper and Lending Club, but could soon encompass additional platforms and additional products. Edmundson and his team leverage data science, machine learning, and human insight to identify the best loans and reduce default rates. “People are inconsistent,” says Edmundson. “Our goal is to figure out what makes borrowers tick — the best way to answer that is by having smart people asking the right questions and then using technology to answer those questions in a meaningful and statistically significant way.” He says that he and his team are focused on doing full re-underwriting of the loans they purchase and understanding risk down to the loan level.  Theorem’s data scientists use rigorous quantitative methods to validate the underwriting of the underlying platforms and to identify any inefficiencies should they exist. “We build loans that have low correlation, that are diversified, and that are going to be steady across the cycle,” says Edmundson.

Edmundson remains excited but also realistic about the future of marketplace lending — “The industry is clearly on the right track, but there will inevitably be a few bumps in the road, which I think will most likely be driven by investors.” He advocates sustainable industry growth and advises the investor community not to overzealously chase yield, citing that it’s important to keep in mind that every reward needs to be taking the context of a risk. ”Ultimately, we are building this ecosystem for the benefit of the borrower. Our responsibility is to give good borrowers access to good credit at good rates, and to do some in a fast and efficient way. The best way we can do that as investors is by allocating capital openly and fairly,” he says.

Contemplating the bigger picture, Edmundson feels that allocators are generally underappreciating marketplace lending right now — “Many of them seem to be focusing on the cycle rather than the trend. Any credit asset is going to have a cycle but the trend is what’s important. Marketplace lending is how people are going to get loans in the future and everyone is going to have an allocation to this like they do today, for example, with bonds.” Even if an allocator isn’t deciding to invest in marketplace lending, Edmundson suggests that it’s important to get ahead of the curve in understanding the asset class and its microstructure. “People who understand this trend are going to do incredibly well — I would tell them, don’t forget to see the forest for the trees.”