About

I studied mathematics and it shaped how I work: define the problem precisely, strip away what does not matter, and judge a solution by whether it holds up outside ideal conditions.

I am currently on the Applied AI Tech team at SK Telecom. Outside work, I explore AI physics for simulation and run an AI project for finance, part of which I maintain as open source.

Selected Work

AI for Science

Preliminaries for Computational Simulation

A bridge from classical simulation tools like FDM and FEM to the kinds of problems neural solvers are actually trying to solve.

Read →

Neural Solvers in Scientific Simulation

A series on how deep learning meets differential equations, from neural ODEs and PINNs to neural operators for scientific simulation.

Read →

AI for Finance

TerraFin

Public open-source core of a broader AI4Finance workflow for data, analytics, interfaces, and agent-ready tooling, with roots in LLM-based sentiment prediction.

Read → GitHub →

Background

Education

Seoul National University

M.S. in Mathematical Sciences — Valedictorian, College of Natural Sciences

B.S. in Mathematical Sciences, Minor in Industrial Engineering

Contact

Questions on a post or ideas for collaboration are both welcome. Best reached by email.

Ki-Ung Song