For the past two years, a small team of former hedge fund managers, quants and astrophysicists have quietly been working on a complex project to bring artificial intelligence in financial markets to the masses.
Now that they’re ready to launch, they’ve been hiring senior people from investment banks to get the product off the ground.
Quant Insight was founded in 2014 by Krishnan Sadasivam, a former partner and macro portfolio manager at Brevan Howard and Mahmood Noorani, who worked at UBS, Bluecrest Capital Management and Citi Capital Advisors. The two worked together at Millennium Capital Partners in their last positions before launching the start-up.
Noorani says that the aim is to arm investors with the quantitative knowledge they need in an environment where computers are increasingly beating human traders. Older methods of valuing securities – such as multiple regression – are no longer adequate in a world where financial markets are influenced by dozens of variables and subject to unpredictably volatility, he says.
“It’s increasingly difficult for discretionary fund managers to gain an edge because financial markets have become so complex over the past ten years, and the amount of available data has expanded,” he says.
Artificial intelligence for everyone
Noorani said that the idea for Quant Insight came in 2010, but not only were they in full-time employment, technology wasn’t advanced enough to support it.
"I wanted to make artificial intelligence in financial markets available for everyone, not just the preserve of the top quant hedge funds out there,” he says. “When we started in 2010, the technology wasn’t there. Now, the cost of processing power has collapsed – we just bought a computer running 100 dual core processors for around £10k. This opens up markets to technology."
The firm is working with Professor Michael Hobson, who is a professor in Astrophysics at the University of Cambridge and an expert in Bayesian statistics and machine learning. Tapping into cosmology expertise might seem a little strange for a financial services firm, but there are parallels.
“There are around 20-30 factors that influence the price of a financial security, and that creates millions of data points to analyse. For an astrophysics expert used to studying the universe, financial markets are child’s play,” he says.
Hiring from investment banks
There are just 16 people working for Quant Insight, but it’s a curious mix. They have a team of programmers and quants working in India to develop the product. Then there’s the academic influence from Cambridge, but the key focus for hiring has been to bring in people with a deep understanding of financial products.
“We need domain expertise, because it’s really important that we understand the drivers behind each asset class we cover,” says Noorani. “We understand this, and the relationships between asset classes. The hedge fund and investment banking experience we have ensures that we our data gives the most valid results.”
In the past month, Ernest van Vredenburch, the former head of EMEA macro sales at Credit Suisse, joined Quant Insight after 18 years at the bank. Huw Roberts, a former senior director in rates sales at Credit Suisse, also joined this year, and in September 2016, Peter Tran joined as VP of sales from BGC Partners.
“It’s been a very challenging time for the macro businesses of most European investment banks, with most undergoing significant structural change to align themselves with the new regulatory framework,” says van Vredrenburch. “After 18 great years at Credit Suisse, I was ready for a different challenge. I have joined as a managing partner to further commercialize Quant Insight.
Hedge fund portfolio managers are changing the way they work. The industry’s high fees are increasingly being questioned in the face of lacklustre concerns and quant funds are suddenly out-performing the rest. At the same time, traditional discretionary fund managers have begun relying on reams of third-party data to try an gain an edge, or are using algorithms in their decision-making processes – the so-called ‘quantimental’ approach.
But Noorani says most third-party data merely scratches the surface on what really influences the price of a security.
“There’s a lot of third-party data out there and fintech firms tapping into this,” he says. “One example is a firm that tracks car insurance policy sales every month to accurately predict movements in auto sales and big auto companies. What we’re saying is that it’s just one of about 30 variables that would affect its price, and we look at them all together.”
Van Vredenburch says that the Quant Insight will be hiring.
“There is significant scope for further sales expansion,” he says. “Right now, we're focused very much on the macro sector across both fixed income and equities types. But QI will also have a key role to play across retail investors.”
Photo: Getty Images