I'm Joonhun Lee from 🇰🇷 and an quant AI researcher
for now.
My interests fall into three categories:
Data, Finance, and Strategy. I am passionate
about solving various problems through data-driven approaches,
particularly in financial tasks.
Owning and leading production-grade decision-making systems for execution, asset allocation, and financial intelligence — expanding from reinforcement learning policies to agentic LLM-based workflows under real-world constraints.
Built quantitative trading infrastructure, including modular backtesting systems and real-time data pipelines, enabling systematic strategy research and decision analysis under realistic execution constraints.
Served as a strategic bridge between product, data, and operations, translating abstract business goals into actionable decision metrics and execution plans.
Proposed a representation learning framework that aligns stack-wise latent features using Sinkhorn divergence. This approach enables the model to learn invariant representations across source domains, ensuring robust forecasting accuracy even when target domain data is inaccessible or highly non-stationary.
Devised a continuous representation learning framework that rethinks masked image modeling through implicit neural representations. Unlike discrete pixel-based methods, MINR achieves strong robustness to various masking strategies and domain shifts while remaining lightweight and parameter-efficient.
Mechanical Watch, Distilled Liquor, Golf & Tennis