Data Context Infrastructure and Next-Generation Data Architecture for AI Agent Reasoning
Session Overview
Companies are overflowing with data. Yet when you ask AI a question, the answer often falls short — why? For AI to make good decisions, it needs more than raw data points; it needs the context that connects them — how this customer relates to that account, how this company connects to that indicator.
The problem is that this context isn't stored anywhere today. It lives scattered across a manager's head, buried inside complex data-extraction queries, or lost somewhere in meeting notes.
Adding context to data and storing it as organizational knowledge — that's a knowledge graph, and it's why the financial industry has recently begun looking to knowledge graphs as the next stage for AI agents. In this session, three practitioners who have each faced this challenge in different fields come together to share their perspectives.
Speakers
Hardy Jeong began his career as a graph data scientist at a graph database company, then worked as a graph data solutions engineer, and now applies graph technology across the semiconductor industry — bringing expertise that spans both software and hardware. He has also contributed to expanding the reach of graph technology through tutorials at domestic academic conferences, and was recently invited to present as Korea's representative speaker at an overseas industry conference (KGC 2026), earning recognition for his expertise in graph technology both in Korea and abroad.
Jaeyoung Jung is a Data Scientist researching and applying AI in fraud detection (FDS) and anti-money laundering (AML) for real-world financial services. At Lambda256, Dunamu's blockchain affiliate, he conducted deep learning research to detect fraudulent transactions and prevent money laundering on the Ethereum blockchain.
At Toss Bank, he has developed an LLM-based agent that streamlines AML report writing and investigation work, and builds and operates machine learning models on the FDS team that detect and block fraudulent accounts and suspicious transactions in real time. He is focused on applying AI to financial risk management to improve operational efficiency and decision accuracy, building a safer, more trustworthy financial environment.
Hyungjoo Lee leads work applying AI/ML in real services and on the ground as Head of the AI Service Center at Kakao Pay Securities. He has a particular interest in structuring data into formats AI can use effectively, and in the preprocessing that makes that possible.
He built experience launching new businesses at SK Communications, Neowiz Games, and MBC, then served as Team Lead of Digital Marketing at Hyundai Card and CPO at Crowdworks. He now leads the AI Service Center at Kakao Pay Securities, delivering real AI outcomes directly to customers and leading the team's work.