About

Welcome to my blog. I am Dongjae Lee, a Ph.D. student in the School of Computing at KAIST and a member of Prosys Lab, advised by Kihong Heo. My research focuses on building trustworthy programming systems by combining formal methods and agentic AI. I am particularly interested in the following directions:

  • Proof-Carrying Spec-Driven Development: Modern coding assistants (e.g., Cursor, Claude Code, and Antigravity) still require human supervision. My goal is to reduce this bottleneck by enabling them to autonomously verify their own code using formal verification techniques.
  • Mixing Natural and Formal Languages: Most vibe-coding prompts and development plans generated by LLMs are written in natural language. However, such descriptions cannot be directly analyzed using formal techniques such as static analysis or type checking. Moreover, natural language is not well-suited to expressing algorithmic control flow, because the scope of loops and conditional statements is often ambiguous. I aim to design a next-generation programming language that blends natural and formal languages, enabling formal analysis while remaining easy to write.
  • Security for Agentic Software: Modern software increasingly incorporates agentic AI, in which systems autonomously execute tasks through prompts and tool-based programs such as Claude Code’s SKILL feature. This paradigm introduces new security vulnerabilities that traditional program analysis cannot capture. My work explores new techniques for analyzing and securing such agentic software systems.

Education

  • M.S. in School of Computing, KAIST, South Korea
  • B.S. in School of Computing, KAIST, South Korea

Publications

2025

Honors & Awards

2020

  • Dean’s List (School of Freshmen, Spring Semester)

2021

  • Dean’s List (College of Engineering, Fall Semester)

2022

  • National Excellence Scholarship

2023

  • National Excellence Scholarship

2024

  • B.S. Summa Cum Laude (GPA: 4.03/4.3)
  • Finalist, Global AI Idea Lab Contest (CODEGATE 2024)

2025

  • Winner, AI Champion Competition (Minister of Science and ICT Award, Team: autophone)
  • Outstanding Teaching Assistant, Fall Semester (CS424 Program Reasoning)

2026

  • Master’s Thesis Excellence Award

Contact