"It goes beyond a simple internship level; it is complete enough to oversee the entire scope of artificial intelligence (AI) development through formal hiring."
On the 22nd in Gwanak District, Seoul, at Seoul National University's Wooseok Economics Hall. Watching the presentations by participants in the advanced big data Fintech AI specialist program (ABS), a smile spread across the face of Ryu Geun-gwan, a professor in the Department of Economics at Seoul National University. Despite a biting cold snap of minus 10 degrees, the building was filled with the heat and tension of the students. Confidence showed in the voices of the trainees on the podium, and the eyes of the faculty and corporate representatives watching were sharp.
That day at Wooseok Economics Hall, the ABS course completion ceremony was held along with a corporate capstone (corporate-linked practical application program) results presentation. The ABS course differs from typical theory-centered education. When corporations provide the thorny problems and raw data they face in the field, students form teams and spend about three months stationed at the company to find solutions in a "hands-on immersion" program.
◇ Report writing 40 minutes → 10 minutes with AI development… Kiwoom says "internship hiring"
In the first presentation, trainee Kim Yeon-su, 27, explained the "AI chatbot" project carried out by the Kiwoom Asset Management investment management team.
"We introduced an AI assistant and developed a tool that helps write reports. It serves as a data control tower and even assists with data search. Based on reports from overseas fund managers (GPs), it also supports writing monthly management reports."
Using the AI assistant, the time typically required to write a report, about 40 minutes, was reduced to around 10 minutes. It also had a strong effect in preventing human error, such as typos.
Kiwoom Asset Management is piloting the AI chatbot and aims to introduce it into operations in March. The five students who joined the project with trainee Kim Yeon-su decided to proceed with an additional internship at Kiwoom Asset Management. Kim said, "Through the internship, we will make sure this project takes root within the company."
After the presentation, the advisory faculty continued with questions. Park Hyun-woo, a professor at the Graduate School of Data Science at Seoul National University, said, "I can sense that the level of capstone projects has risen considerably," and asked, "What do you see as the success factor?" Kim answered, "It was decisive that we exchanged daily feedback with practitioners during the Natural Language Processing (NLP) process."
Park Woo-jin, a professor in the Department of Industrial Engineering at Seoul National University, pushed the task a step further. "It is close to an automation system now, but it can be extended into decision-making areas, such as fund performance evaluation," Park said. "It is also worth considering supplementing the subjective evaluations of professional institutions with statistics and AI analysis."
◇ Tracking tariff shocks with an ultra-large "400,000×400,000" matrix
Analysis also continued to quantitatively capture shocks across industries. The team collaborating with NICE Information Service analyzed the ripple effects of U.S. tariff measures across domestic industries. Based on the Leontief input-output model of the inter-industry table, they used AI to build an ultra-large analysis matrix of "400,000×400,000." They numerically demonstrated how the sales changes of particular corporations propagate along transaction relationships.
Advisory professors fired off questions: "Can it be extended to other risk factors such as exchange rates, raw material prices, and freight expense?" "How can it be used in the industrial field?" "Can you also show how to improve the network to minimize ripple effects?" The presenting team explained, "It is possible by adjusting variables," and added, "We are also considering improvement models using Machine Learning and Deep Learning, but we have not yet reached the implementation stage."
The team working with the Ministry of Data and Statistics (MODS) unveiled an "automatic regional economic trends press release generation system." However, the team faced difficulties as a member dropped out mid-course, and it proceeded as a one-person project. Even so, they implemented a program that automatically writes press releases containing numerical interpretation and explanation based on regional economic statistical tables.
The Hangul word-processing program widely used in public institutions cannot be directly edited by AI, revealing a limitation that requires copy-and-paste steps. In light of the current policy stance that advocates an "AI government," there were notes that structural improvements to the work environment in the public institutional sector are needed.
In addition, six teams presented capstone projects carried out in collaboration with various corporations and institutions.
Professor Ryu Geun-gwan said, "The Seoul National University ABS course is an AI education program combined with hands-on work," and added, "Students themselves carry out corporations' AI innovation projects, and based on this, they even move on to patent applications."
Ryu said, "The completion ceremony is also a venue where corporate management taking the big data AI CEO (ABC) course and students seeking professional positions meet," and added, "It is a place where education, hiring, and industry meet on the ground."