22 March 2022 – A team of four Master of Science in Business Analytics (MSBA) graduates won first place at the AIxImpact Case Competition (Finance stream) in February this year. The case competition was held as part of the AIxImpact Conference, a Southeast Asia-focused analytics conference hosted by QuantumBlack, a McKinsey Company, in early February.
The competition is designed to get students across Southeast Asia to solve real-life problems with AI. Participants were required to work on one out of three problem statements, each belonging to a theme: finance, health, or e-commerce.
Gino Martelli Tiu, Rachel Sng, Widya Salim and Xhoni Shollaj, all recent MSBA graduates, won the first place for their solution, a recommender system that tackles the issue of making wealth management accessible to the middle class in Singapore. This accounts for over 4 million adults in Singapore alone.
The finance-themed problem statement they chose to tackle required them to use data analytics to enhance the customers’ banking experience while respecting customers’ privacy.
Structured as a SAAS integrator, the team’s solution was a personalised plug and play recommendation system called the Money Tree Integrator, built off a federated learning framework.
The service essentially allows banks easy access to best-in-class recommender solutions from day one. It removes the initial dependency on financial planners, and matches customers to products based on perceived needs, while respecting the need for privacy in a way that is scalable and quick to deploy.
Carefully curated algorithms also solve the problems of cold-start and incomplete information, which are the bane of many recommender systems in industry. The cold-start problem refers to the issue of items in a catalogue having either none, or very little interactions.
“Tackling a topic that was socially relevant and technically challenging was very fulfilling. This opportunity allowed us to see first-hand how the stuff we’re learning can eliminate traditional barriers and bring positive change at scale,” said Tiu.
Added Sng: “We feel very honoured to be able to represent NUS MSBA in this regional competition. It was great that our proposed solution was recognised to be both technically strong as well as addressing a real business need.”
The team faced several challenges when it came to developing the recommender system.
“We hadn’t learnt about recommender systems yet in the course, but we researched extensively to ensure that the solution was technically strong, as well as feasible. We also had to place the proposed system in the context of the financial products, and include solutions to ensure data privacy, as it is a more pressing concern for the finance industry compared to typical consumer applications, like ecommerce or online entertainment,” explained Salim on behalf of the team.
“For us, it meant a lot to be able to participate in such a prestigious competition, and to beat numerous brilliant and innovative teams in the process. We worked really hard to bring forth a solid solution, and achieving first place was exhilarating, especially against other ingenious innovations presented in the competitions,” added Shollaj.
The team won books from QuantumBlack, a three-month mentorship programme with assigned mentors from QuantumBlack, well as exclusive trainings and technical sharing sessions.