Experience

Click for complete list.

Work Experience

Nota AI

AI Research Engineer

Jul 2024 - Current

Developing technology to deliver the value of AI to more people in various environments.

  • Working on VLM for on-device applications to solve real-world problems.

Deargen Inc.

AI Scientist

Sep 2022 - Jun 2024

Worked on developing AI models in a field of drug discovery for a better world.

  • Developed a representation learning methodology capable of handling various scales of molecules, from small molecules to proteins, in a universal embedding space.
  • Developed a model for controllable molecular generation, depending on protein conditions, based on the diffusion model.
  • Discovered an imbalance in the process of models performing molecular interaction prediction utilizing features from different modalities and resolved it through a multimodal-mixing method.

Project Experience

Personal Projects

Deepest: SNU Deep Learning Society

NCIA Lab

Growth Hackers: Business Data Analysts

Greenhouse Gas and Energy Management Center

Student Worker

Jan 2020 - Aug 2020
  • Managed and visualized Seoul national university's greenhouse gas and energy data with R programming.
  • Improved data management workflow and established visual baselines for tracking energy consumption trends and identifying anomalies in buildings.
  • Realized the importance of experiencing and handling actual real-world data.

Research Experience

Current Work

Research Topics in Interest
  • Better VLM design
    • Integrating better vision encoder designed to handle both image and video efficiently.
  • DL for Simulation
    • Accelerating the process of simluation-based science and engineering via DL.

Past Work

Researched Topics
  • DL for Drug Discovery
    Sep 2022 - Jun 2024
    • Meaningful performance estimation method for the drug discovery domain.
    • Meaningful representation learning to handle various scale of molecules universally.

During M.S.

Sep 2020 - Aug 2022
Researched Topics
  • Neural Tangent Kernels (NTKs) and DL theories
    • NTK emerges in the infinite-width limit of a NN, suggesting that characteristics of DL model can be simplified to a linearized model under certain conditions.
    • The trend towards larger NNs has motivated the study of initialization and training at large network width, leading to the infinite-width assumption for studying their dynamics.
    • Thus, from a practical viewpoint, modern overparametrized DL models can be analyzed via NTK with infinite-width limit as feature dimension grows more and more.

  • Generative Models (Diffusion and Schrödinger Bridge models)
    • The recently proposed diffusion models show excellent performance in various domains and tasks.
    • Schrödinger Bridge (SB) models, which are based on an entropy-regularized optimal transport problem, can be interpreted as an extension of the diffusion models.

During B.S.

Mar 2016 - Aug 2020
Undergraduate Research Internship
  • Topic : Theoretical Background on the Convergence of Optimizers Used in Deep Learning
    Mar 2020 - Aug 2020

  • Topic : Exploring the Mathematical Background Related to Machine Learning(Reinforcement Learning)
    Jun 2019 - Aug 2019

Course Experience

During M.S.

Sep 2020 - Aug 2022
Key Courses
  • Machine Learning for Visual Understanding
    • Team Project: Image Deblurring with Generative Diffusion Process
  • Practical Application Research of IoT·AI·Big Data
    • Personal Project: Various GANs and its Application(Face Aging)
  • Mathematical Algorithm
  • Numerical Analysis

During B.S.

Mar 2016 - Aug 2020
Key Courses
  • Mathematical Modeling and Computational Experiments
    • Team Project: Text Detection based on YOLO v3
  • Combinatorial Optimization
  • Linear and Non-linear Optimization
  • Operation Research

More Topics

Never ends.
Textbooks
  • [Sumio Watanabe] Algebraic Geometry and Statistical Learning Theory
  • [Cuturi] Computational Optimal Transport
  • [Kevin Murphy] Probabilistic Machine Learning
  • [de Berg] Computational Geometry
  • [S. Boyd] Convex optimization
Open Courses
  • CS224W: Machine Learning with Graphs by Stanford University
  • CS329S: Machine Learning Systems Design by Stanford University
  • CS231n: Convolutional Neural Networks for Visual Recognition by Stanford University

I'm eager to learn new things constantly.

Teaching Experience (TA)
  • Elementary Mathematical Analysis
    Spring 2022
  • Linear Algebra
    Spring 2022 / Fall 2021
  • Calculus
    Fall 2021 / Spring 2021 / Fall 2020