GEON CHOI.
Systems reveal
their own shortcuts.
Rather than treating systems as black boxes, I look for the constraints, regularities, and asymmetries that reveal a more precise way to design them.
Structure
before scale.
More computation is not always the answer. I study how systems behave—where complexity accumulates, which regularities emerge, and how those structures can guide deliberate choices about computation and information.
THE QUESTIONHow can a deeper understanding of a system lead to a better way of designing it?
- 01Observe
Study concrete system behavior instead of relying on generic assumptions.
- 02Model
Expose the relationships and asymmetries hidden inside the problem.
- 03Constrain
Make performance, resource, and reliability boundaries explicit.
- 04Optimize
Use structure to remove unnecessary search and computation.
PROFILE / 2026
I am a Ph.D. student in AIDAS Lab at Seoul National University. My research focuses on AI systems, with current work in vector databases and approximate nearest-neighbor search.
My broader interests include adaptive ML infrastructure, software reliability, information retrieval, and embodied AI. I received a B.S. in Computer Science and Engineering from Sogang University.
Evidence,
not ornament.
ACM SIGMOD 2027
Exploiting Structural Properties for Efficient Constraint-Aware HNSW Hyperparameter Tuning
Geon Choi, Hoeun Lee, Jaeyoung Do
NEURIPS 2026
Dynin-Robotics: Omnimodal Unified Diffusion Vision-Language-Action Model
Hoeun Lee, Jaeik Kim, Jusang Oh, Jinhyeok Kim, Geon Choi, Hyeonggeun Kim, Jaeyoung Do
04 / CONTACT · RESEARCH INQUIRIES