Quantitative Finance & NLP Enthusiast
Korea University (Seoul, Korea)
- B.A. in Linguistics & International Studies
- Cumulative GPA: 4.09/4.5
- Major GPA: 4.30/4.50
- Global Leadership Program Japan @ Doshisha University (Jan 2021)
- Excellence Awards ×4 in Linguistics & International Studies
- NLP-driven portfolio optimization using sentiment scores
- Option-data signals for US equity trading
- Volatility surface construction (SABR & extensions)
- Fine-tuning financial language models (LoRA / QLoRA)
- Retrieval-Augmented Generation
| ML / DL |
|
| Web |
|
| Container & Cloud |
|
| ETC. |
|
- Enhanced stock price prediction accuracy by integrating Aspect-Based Sentiment Analysis (ABSA) into the conventional finBERT model
- Curated and augmented a domain-specific dataset from online stock-related news, with labeling supported by a Large Language Model (LLM)
- Team: Prometheus AI Club
- Result: Finalist; top scores in Financial Markets (59.6) & Lonform (2nd)
- Pipeline Highlights:
- Crawled & cleaned textbooks, CFA/Q-bank, K-IFRS → OCR + Markdown tables
- Built high-quality MCQA dataset → Instruction-tuned LLM (Gemma-3 & Qwen variants)
- Retrieval-augmented QA evaluation with fold-wise validation
- Press Coverage: 연합뉴스 기사
- Modeled analyst, portfolio-manager, and trader personas with dedicated GPT agents
- Built RAG database and real-time data fetcher to react to market changes instantly
- Monitored BTC price lag between Binance & Upbit
- Executed arbitrage orders when deviation thresholds exceeded
- Clustered Upbit white-papers into 7 groups via CryptoBERT encoder
- Transfer-learned on new listings for low-data prediction; backtested risk-adjusted returns
Korean Standards Association
- Result: Excellence Award
- Compared public vs. private sector digitalization costs & benefits
- Conducted inter-governmental data research, gap analysis, and strategy recommendations
KOPOMS & Hyundai Motors
- Result: Grand Prize
- Analyzed delivery-related social costs—including traffic congestion and safety impacts—using public big data
- Designed a data-driven architecture for robotics-based mobility solutions to mitigate social externalities
- Developed quantitative derivatives models (Black–Scholes derivations, jump-diffusion)
- Researched options-greeks-based trading strategies
- Implemented core ML demos in vanilla JS & Python for government AI textbooks
- Collaborated on large-scale educational content pipelines
- Explored mathematical foundations for AI/ML, including linear algebra, statistics, and deep learning.
- Practiced large-scale collaborative project development
- Blog: sit-in-a-row.vercel.app