The University of Melbourne
EducationPh.D. Student · Faculty of Engineering and Information Technology
Supervisors: Negin Yousefpour and Guillermo A. Narsilio
About Me
I am Changjian Zhou, a Ph.D. student in the Faculty of Engineering and Information Technology at the University of Melbourne.
My research interests lie in physical AI, scientific machine learning, and computational mechanics. I aim to build reliable, physics-aware models that learn from simulations and adapt to real-world data.
Education & Experience
iFLYTEK Co., Ltd.
IndustryFull-timeMachine Learning Engineer · LLM
Core Development Platform · Adaptive LLM reasoning with supervised fine-tuning and reinforcement learning.
Tsinghua University
ResearchFull-timeResearch Assistant
Computational fluid dynamics of bone cement flow, in collaboration with medical and engineering teams.
Shanghai Jiao Tong University
EducationM.Eng. · Civil and Hydraulic Engineering
Supervisor: Guanlin Ye
Hefei University of Technology
EducationB.Eng. · Hydraulic and Hydropower Engineering
Research Interests
- Physical AI & world models. Learned simulators for physical interactions and contact dynamics.
- Scientific machine learning. Physics-informed learning and sim-to-real adaptation.
- Computational mechanics. Granular dynamics, constitutive modelling, and inverse problems.
Publications
Google Scholar ↗7 publications · Newest first
A weak-form-based 1D numerical model for guided wave dispersion in subsea buried pipes embedded in saturated poroelastic soil
Ocean Engineering, 363, 126622
Stochastic Polynomial Surrogate Models for Uncertainty Quantification of Offshore Plate Anchor Capacity
Geotechnical and Geological Engineering, 44, 305
Parameters sensitivity and identification in the Shanghai Model: A numerical analysis for deep excavation
SSRN 5179194 Preprint
Research software & open-source contributions
PhysGuard Research code
Fisher-guided gradient projection for adapting neural PDE surrogates from simulations to experimental data while preserving learned physical structure.
TRACE Research code
A graph-network simulator that preserves contact history on edges and uses a physics-structured decoder for granular dynamics.
Constitutive Parameter Identification Research code
Machine learning and search-based calibration of soil constitutive models, with companion tools for automated Abaqus parameter sweeps.
Auto Abaqus Agent Research Engineering tools
An AI agent for Abaqus workflows, including convergence diagnosis, mesh assessment, input generation, and output analysis.
I also contribute to DeepCopilot (a coding assistant for VS Code) and StructureClaw (an AI-assisted structural engineering workspace).
Selected Honors
- Outstanding Graduate Honor, Shanghai Jiao Tong University (2024).
- Yang-Yuqiu Scholarship, Shanghai Jiao Tong University (2023).
- Second Prize, China Postgraduate Mathematical Contest in Modeling (2023).
- Distinguished Undergraduate Thesis Award, Hefei University of Technology (2019).
- Second Prize, China Undergraduate Hydraulic Innovation Design Competition (2017).
Research Collaborations
Institutions represented in my coauthored research.
Melbourne, Australia
- The University of Melbourne
Singapore, Singapore
- National University of Singapore
Oslo, Norway
- Norwegian Geotechnical Institute
Hefei, China
- iFLYTEK
Beijing, China
- Tsinghua University
Shanghai, China
- Shanghai Jiao Tong University
- East China Architecture Design & Research Institute (ECADI)
- Shanghai Engineering Research Center of Safety Control for Facilities Adjacent to Deep Excavations
Chengdu, China
- Southwest Jiaotong University
Hover, focus or tap a location · Sources · Natural Earth
I welcome conversations about scientific machine learning, computational mechanics, and research software. The best way to reach me is by email. Get in touch ↗