Jaehwan Jeon

Jaehwan Jeon

Jaehwan Jeon is a Postdoctoral Fellow in Civil and Mineral Engineering at the University of Toronto. His research tackles the two bottlenecks of structural safety prediction — computation and expertise — scaling both with AI while making that AI trustworthy.

Jaehwan Jeon (전재환), PhD

About

I am a Postdoctoral Fellow in Civil and Mineral Engineering at the University of Toronto. Structures are never crash-tested — a building is built once, and people move in the next day. Its safety exists only as a prediction. My research tackles the two bottlenecks of that prediction, computation and expertise, scaling both with AI while making that AI trustworthy: physics-informed deep learning to overcome the computational cost of nonlinear response analysis, uncertainty quantification to keep predictions honest, and multi-agent LLM frameworks to lower the expertise barrier to structural reliability analysis.

Research interests

  • Nonlinear structural response prediction with physics-informed deep learning — fast, physics-consistent surrogates for probabilistic safety assessment
  • Uncertainty quantification and decomposition for probabilistic surrogate models of structural behavior — using Bayesian neural networks
  • Automating structural reliability analysis with trustworthy multi-agent LLM frameworks
  • Reliability and resilience of structural and energy infrastructure systems under multi-hazard risks (ongoing)

Research highlights

  • Automating structural reliability analysis with a multi-agent large language model framework
    Preprint
  • Neural network–augmented physics models for scalable dynamic MDOF systems
    Paper
  • Unified hysteresis modeling via physics-based deep learning and data augmentation
    Paper · Code
  • Uncertainty decomposition in probabilistic surrogate ensembles using Bayesian neural networks
    Paper

Contact