APU’s Milodino team, formed by Actuarial Studies specialism in FinTech students Deaxell Maverick Suryanto, Neela Bala Subramaniam, Wendy Mah Yun Nee, and Patrick Ian Dass, developed a sophisticated machine learning framework that automates customer complaint classification, supporting Bank Negara Malaysia’s (BNM) efforts to improve complaint handling in financial services, and secured 1st Runner-Up at the UMACT Hackathon 2025.

A team of bright Actuarial Studies with a specialism in Financial Technology (FinTech) students from Asia Pacific University of Technology & Innovation (APU) has secured First Runner-Up at the Universiti Malaya Actuarial & Financial Mathematics Society (UMACT) Hackathon 2025.

Team Milodino, comprising Deaxell Maverick Suryanto, Neela Bala Subramaniam, Wendy Mah Yun Nee, and Patrick Ian Dass, developed a sophisticated machine learning framework in response to Bank Negara Malaysia’s (BNM) drive to streamline complaint resolution across the financial services sector.

“Our innovative solution aims to automate the classification of customer complaints, a crucial step in improving efficiency through a system designed to automate workloads, drastically cut down resolution times, and ultimately elevate the customer experience,” explained Deaxell, representing Team Milodino. 

The team also expressed immense gratitude for the invaluable guidance provided by their mentor, Mr Lee Chee Nian, whose expertise proved instrumental in their journey.

Utilising advanced text mining, natural language processing (NLP), and sentiment analysis, the team meticulously sifted through a vast database of consumer grievances, allowed them to unearth vital insights and develop a model capable of accurately tagging and categorising issues. 

Presenting their findings and recommendations to a panel of seasoned industry judges was no small feat, but it proved to be an incredibly rewarding part of the experience for the team.

The opportunity to address a real-world industry problem with significant implications for consumer financial protection energised the team.

The sheer potential of their work to revolutionise how banks and insurers handle complaints spurred them on. 

Their shared passion for data science, coupled with the pressure of applying their skills in a high-stakes environment, made the hackathon an invaluable learning and growth opportunity.

Milodino proudly reflected on the team’s journey, describing it as “Overwhelming but highly meaningful.” 

Despite the considerable challenge of preparing during their final exam week, the team remained remarkably cohesive and determined.

Deaxell expressed profound gratitude for each team member’s commitment, unwavering resolve, and excellent cooperation, attributing their collective spirit to making the experience both memorable and successful.

“It was a real-world problem-solving exercise, blending data, regulation, and technology. 

“As this was only our second hackathon together as a team, it felt great to win. Our success was down to effective teamwork, clever model choices like TF-IDF + logistic regression, and a deep understanding of the business problem, especially BNM’s compliance requirements. 

“We also spent a significant amount of time perfecting our presentation, ensuring our solution was not only correct but also understandable, scalable, and viable in the real world.”

According to Deaxell, the hackathon taught them how to apply machine learning to practical business scenarios, strike a balance between technical accuracy and regulatory demands, and communicate complex ideas clearly to a diverse audience. 

He continued by saying that APU “Forced us to grow both technically and professionally.”

“It was challenging when we spent the entire day sitting together and trying to come up with the best solution, especially as it was so close to our final exams,” he shared.

“In identifying various modelling abilities and choosing the best, we gained a clear understanding and knowledge from this hackathon.”

For team Milodino, this hackathon undeniably served as an extension of their classroom learning, pushing them far beyond their coursework in modelling, analytics, and problem-solving.

“Real-world data is not like what we get to study; the data is always raw, complex, and much more challenging,” the team stated unanimously, added “For us, the development after this hackathon was the true success.”

They are also looking forward to competing in future events, as they have their sights on the CAS Hackathon that will soon be taking place.