Research Interests

Machine Learning, dealing with manifolds and optimization for models in industrial or medical domains such as continuous data and tabular data, understanding model outputs for generalization. Convex optimization and solving generalization problems (grokking) through mathematical methods such as Taylor series, Fourier or Laplace transformation. Finding the learning mechanism behind memorization and understanding the mechanisms of Neural Fields.

Education

Mar 2020 — Present
Sungkyunkwan University, Suwon, Korea
BS., Computer Science and Engineering
GPA: 3.90/4.50

Mar 2016 — Mar 2019
High School Diploma, Yeongdong High School, Seoul, Korea

Professional Experience

AI Researcher, The Catholic University of Korea St. Vincent’s Hospital

Sep 2024 — Feb 2025

  • Implementing machine learning algorithms to create protocol
  • Feature engineering patient data and handling imbalanced datasets with methods such as K-NN imputer, t-SNE, Tomek-links
  • Extracting meaningful vectors for categorical features using Sent2Vec
  • Using Imbalanced-XGBoost and handling hyperparameters to successfully achieve desired f1-scores

Undergraduate Research Student, iisLab SKKU

Jun 2024 — Aug 2024

  • Attended as an undergraduate research student in iisLab led by Prof. Ji-hyung Lee
  • Participated in seminars for subjects such as Computer Vision, Natural Language Processing
  • Presented in Machine Learning Seminars for subjects such as Unsupervised Learning, Optimization methods, Recurrent Neural Networks

Member, Artificial Intelligence College Conference, TNT SKKU

Mar 2024 — Jun 2024

  • Participated in AI College Conference Test aNd Train for weekly studies
  • Prepared and presented papers about self-RAGs and Mistral attention mechanisms
  • Joined and conducted studies about the basics of Reinforcement Learning

Industry-Academia Collaboration Software Intern, SOYNET

Jun 2023 — Aug 2023, Seong Nam

  • Participated in a year-long Industry-Academia Collaboration Project between SKKU & Soynet
  • Learned the basics theories of machine learning & coding skills
  • Worked at the company as a Summer Internship for 2 months
  • Used PyTorch, CUDA, C++ for AI Inference acceleration and model optimization
  • KIST University Person RE-ID model Optimization Applied to National Safety SQI Soft Task

Projects & Awards

20th Korean Economic Securities Derivatives Competition - 2nd Place

Jul 2024 — Feb 2025, Seoul

  • Worked on designing an ETF for IPO stocks
  • Used Gaussian-Mixture-Model and t-SNE, Weighted K-NN for finding the characteristics of newly IPO companies with Original KOSPI & KOSDAQ market prices
  • For implementing Portfolio optimization used convex optimization methods such as Disciplined Convex Programming and SLSQ Programming
  • Participated in the competition as a Research Engineer
  • Received Korea Exchange Chairman’s Award

LLM-based Recommendation System Health care Startup, iKooB

Jul 2024 — Jan 2025, Gangnam, Seoul

  • Started as a SKKU Research Project for “Exploration of AI model architecture based on transfer learning Development of model data management platform”
  • Collaborated with iKooB, experimented with various methods such as transfer learning on LLMs and Graph Attention Networks (GAT) for Implementing Recommendation system
  • Used a finetuned Llama-3 based 8B model on medical domains, implemented TF-IDF method and Chain-of-Thought (COT), Prompt tuning for Prototype model
  • Implementation stack: vllm, langchain, ngrok, flask
  • Received Engineering Innovation Award from SungKyunKwan University

Winter Global Capstone Design Project, AIIT & SKKU

Dec 2023 — Jan 2024, Tokyo, Japan

  • Started a project Proposing a Mobility cloud infrastructure solution for the FSD era
  • Collaboration with Tokyo AIIT University and participation in seminars and presentations
  • Proposed an effective construction scenario for signaling system cloud infrastructure for mobility
  • Scenario proposal for integrated cloud and AI-based solutions

NH Investment & Securities Big Data Competition Finals

Sep 2023 — Dec 2023, Seoul

  • Designed and trained a Transformer-based correlation analysis model between stocks (Fin2Vec)
  • Used 1D-CNN, GRU for stock price data compression
  • Used Transformer Encoder architecture for correlation analysis and implemented knowledge distillation method by teacher-forcing

Development of SNS platform exclusively for college students

Jun 2023 — Dec 2023, SKKU

  • With support from Sungkyunkwan University Startup Support Group started a SNS app development project: (CLIPPED)
  • Purpose of launching a Startup and Conducted actual testing
  • Participated as a Project Manager and Frontend Developer
  • Implementation stack: Flutter, AWS, next.js

Skills

Programming

Language/Technology Proficiency Details
Python ★★★ Various machine learning & processing libraries such as PyTorch, scikit-learn, Numpy, cupy, cxypy
C, C++ ★★ Conducted many system-architecture college projects such as XV6, Spike-simulator, ns3, mini-shell
CUDA Conducted acceleration tasks to PyTorch functions such as im2col, and XGBoost Minimal Variance Sampling methods
Next, Node.js Developed Web Projects using Next.js and Node.js with tailwind-css
Flutter Developed front-end app for the SNS platform

Languages

  • Korean (Native)
  • English (Fluent) (TOEIC: 930)

Publications & Research

Coming soon

Contact

Feel free to reach out to me at jaheon555@g.skku.edu or connect with me on LinkedIn.