Nicholas Venuti of Morgan Stanley's Machine Learning Research Group turned his passion for data science into a career as a bank technologist.
In college, Nicholas Venuti spent years researching ways to extract oil from coffee grounds to make biodiesel fuel. After he graduated and began working at an environmental engineering consultancy, he discovered his true passion: data science.
At the heart of his enthusiasm was machine learning, a type of artificial intelligence (AI) that allows computers to "learn" by recognizing patterns and making inferences based on enormous sets of data.
Today, as a research scientist in Morgan Stanley’s Machine Learning Research (MLR) Group, Venuti is figuring out how to harness the raw power of data and turn it into useful insights for traders, financial advisors and investment bankers.
“In financial services, the problems are enormously complex, and they’re changing all the time," says Venuti, who was the first data scientist to join the MLR team in 2016. “I always find the work both fun and mentally challenging.”
The MLR, with operations in Montreal, New York and London, is an internal body that helps enable and scale the adoption of AI and machine learning across Morgan Stanley by providing tools, consulting, and research services to business units and their IT teams.
As part of the group, Venuti works with every corner of Morgan Stanley's operations, from the Fixed Income and Investment Management professionals who deal in complex securities to the financial advisors who work with Wealth Management clients.
The complexity and fast-moving pace of data science attracted Venuti to the field. "It's like trying to solve a big puzzle," he says. "You're constantly looking for relationships and tendencies that help explain the problem you're trying to solve.”
Wall Street is a natural place for a data scientist, Venuti says, because investment banks have such enormous volumes of trading data, as well as a continuously growing stream of external information coming from news, social media, analyst reports, regulatory disclosures and elsewhere.
"Every day the environment is changing, and as such, the systems we have to build have to be able to adapt to those changes," he said. "We are competing with other firms all trying to do the same thing, so whenever, you solve a problem, you need to move on to the next one to make sure you stay ahead of the game."
A Technologist Charts a Career Path to a Bank
A native of western New York, Venuti studied biomolecular chemical engineering as an undergraduate at North Carolina State University, then worked at an environmental engineering firm, where he built systems that analyzed data on oil and natural-gas projects.
“During that time, I developed a knack for solving our company’s data problems and building systems to produce insights from these different data sources,” he says. “I found myself wanting to perform more robust analyses and, in turn, started taking online courses to add to my skill set.” Those classes deepened his interest in data science and prompted his decision to attend graduate school.
At the University of Virginia, where he earned a master's degree in data science, he worked on a team that helped the school’s Religious Studies department synthesize the meaning of words in religious texts.
Upon graduation, he was hired by Morgan Stanley to work on AI, a field at the forefront of technology. Now, it is more important than ever. "AI, and its potential in the business and technology world, has been flipped on its head in the last seven years,” Venuti says. “People who were hesitant in 2016 are now some of our largest supporters. What were pipe dreams then are standard practice now, and bleeding edge ideas back then are already starting to become obsolete," Venuti says.
Opportunities to Work on Meaningful Projects
One of Venuti’s favorite areas of research has been developing bespoke deep learning architectures for pricing illiquid instruments in the fixed income space. The schema learns how an illiquid instrument’s price changes based on changes in related, more liquid instruments, and allows fixed income traders to better infer prices for these hard-to-price instruments.
This method was first implemented with specified pools in the mortgage-backed security market and reduced price error by 25% in the primary market and 50% in the secondary market. With its early success, the team plans to expand it to other illiquid securities.
Finding elegant ways to deploy technology is just one goal of Venuti and his colleagues. Another one is demystifying the work they do so that everyone understands just how much it can enhance their roles. “Some people hear the term ‘artificial intelligence’ and assume that existing systems will be completely replaced or that human expertise will be supplanted,” Venuti says. “But AI will compliment human expertise, not replace it. AI can help inform the decision-making process, but we still ultimately need the human touch to make the wisest and most sophisticated choices.”
