Korea Advanced Institute of Science and Technology
– PresentB.S. in AI Computing, double major in Physics, minor in Mathematical Sciences. Second-year undergraduate; began studying machine learning independently in middle school (2021).
Daehan Won
Student Representative, Dept. of AI Computing, KAIST
Intern, MLML Lab
Research interests
I am broadly interested in the foundations of general intelligence — not as an engineering target, but as a scientific question about what kinds of systems can learn to think.
My curiosity centers on three interconnected threads.
First, how learning itself should work: I find conventional gradient-based fitting unsatisfying as a theory of intelligence, and am drawn to alternatives inspired by biological learning — spike-timing, unsupervised structure discovery, and systems that prioritize what to learn rather than passively absorbing all data equally.
Second, representation and memory: I believe the way information is stored and composed matters as much as the learning algorithm itself — particularly ideas around diagram-like or graph-structured memory, and whether agents could develop their own internal representational languages.
Third, the boundary between information systems and biological minds: questions like whether a system taught a model of the world rather than answers to questions could exhibit something resembling understanding — and what it would even mean for it to do so.
These interests connect naturally to world models, neurosymbolic reasoning, next-generation architectures beyond the transformer, and embodied or agent AI.
Education
Incheon Jinsan Science High School
–Early graduation. Self-study and coursework beyond the curriculum.
Experience
Undergraduate Teaching Assistant
–- Directly recruited by Prof. Hong as a second-year undergraduate to join the course development team; one of the few sophomores selected for a TA role involving curriculum design.
- Contributed to the Agentic AI module: primary TA for one assignment (LLM-based agents, prompting, LangGraph) and part of the development of the module’s other assignments.
- Co-authored midterm exam questions; responsible for problem design, solution writing, and rubric specification for the Agentic AI section.
- Held office hours and Q&A sessions, explaining concepts such as search, probabilistic reasoning, and neural networks to students in both Korean and English.
President, Include
–- Led a society of 45 members, overseeing seminar scheduling, project teams, and external collaborations.
- Competed as a team in AI/ML competitions such as DACON and Kaggle, working through full ML pipelines from data preprocessing to model evaluation.
- Active in the club’s paper-reading seminar from first year through the presidency; presented 10 papers on computer vision, NLP, explainable AI, and quantum AI.
Google Gemini Student Ambassador
–- Led the team’s ambassador activities on campus; organized and hosted a Google AI seminar for KAIST students covering Gemini and related AI tools.
- Managed festival booth operations and campus-wide promotion of Google AI initiatives, coordinating outreach and event logistics across the team.