Curriculum Vitae Van Hai Nguyen 📍 Austin, TX | 📞 512-461-2988 | ✉️ haivnguyen2021@gmail.com | 💼 LinkedIn Education PhD in Aerospace Engineering, GPA: 3.9/4.0 Dec 2025 (Expected) The University of Texas at Austin Dissertation: "Model-Constrained Machine Learning Approach for Forward and Inverse Problems" Honor: M. J. Thompson Endowed Presidential Graduate Scholarship (2023) for great research performance Master of Science in Civil and Industrial Construction Engineering Jun 2017 Ho Chi Minh City University of Technology Bachelor Engineering in Civil and Industrial Construction Engineering Jun 2015 Ho Chi Minh City University of Technology Skills Programming Language: Advanced Python (JAX, PyTorch, TensorFlow), MATLAB, Julia, MPI computing Software & Tools: Docker, Git, bash, Paraview, Firedrake, LaTeX, Microsoft Office, MAPLE, Flask, Kubernetes, ANSYS, AutoCad Work Experience Lawrence Livermore National Laboratory, Computing May 2025 - Aug 2025 Computational Science Intern Livermore, CA, USA Project: Physics-based generative AI models for spinodal decomposition from stochastic differential equations (SDE) Developed physics-based generative AI models for accelerating the simulation from 36.67 hours to 30 minutes Performed training models with highly noisy low-dimensional SDE data, generalized well to high-dimensional simulations Performed comprehensive analysis of the stochastic effects on accelerating the spinodal decomposition process Prepared the research manuscript to submit a peer-reviewed journal (On preparation) Chevron Cooperation, Subsurface Innovation Lab May 2024 - Aug 2024 Subsurface Modeling Intern Houston, TX, USA Project: Physics-informed implicit neural representation for geomodelling and geophysics inversion Filed a U.S patent on inventing a unified machine-learning approach for efficiently solving problems in geomodeling and geophysics inversion. It was the first time that an internship project at Chevron returned a U.S. patent Performed data parallelism training on multiple GPU clusters with JAX and enabled high-resolution complex geology objects Accelerated 40× computation time, and saved 60× computation memory compared to legacy grid-based code Improved significantly the performance of FWI with complex salt by removing the need for initial models Published the research work in a peer-reviewed conference Chevron Cooperation, Computation Reservoir Geophysics R&D May 2023 - Aug 2023 Earth Science Intern Houston, TX, USA Project: Physics-informed machine learning for elastic full waveform inversion (FWI) Inverted successfully the subsurface earth models for field data without initial models, for the first time, in academia and industry Reduced drastically from months or more to a few days the cycle time for imaging inversion than the conventional FWI approach Accelerated 57× computation time for solving wave equations, accelerated 23× training time, and saved 35% GPU memory Parallelized training on 8 of A100-GPUs, thus scaling for large-scale 2D&3D synthetic and field data problems Published the research work in a peer-reviewed conference JFE Steel Corporation, Steel Research Laboratory Jul 2018 - Sep 2018 Steel Foundation Internship Nagoya, Japan Project: Research on steel pipe pile foundation design and geological ground survey Carried out a critical comparative analysis of Vietnamese and Japanese foundation design standards, establishing the groundwork for Vietnam's first steel foundation design standard Led a geological ground survey in Vietnam to successfully evaluate the applicability of JFE steel pipe pile foundations Research Experience The University of Texas at Austin Jan 2021 - Present Graduate Research Assistant Austin, TX, USA Project: Model-constrained machine learning frameworks for simulating time-dependent PDEs with discontinuities Designed neural network surrogate models for 10× faster time-dependent PDEs simulations with discontinuities Generalized neural networks for unseen scenarios, including discretization, boundary conditions, geometry, and parameters Parallelized the training models and differentiable numerical PDEs simulations on a GPU cluster of 128 GPUs Project: Model-constrained autoencoder machine learning frameworks for solving PDE-constrained inverse problems Learned inverse solver surrogate models with one arbitrary training sample, thus reducing drastically the training data requirement Accelerated computation time 25,000× faster than the classical Tikhonov framework while achieving the same accuracy level Project: Redesigning Transformer architecture for simulating time-dependent PDEs and forecasting time-series data Redesigned the transformer architecture via the perspective numerical methods for PDEs Achieved a higher-order convergence rate than the vanilla transformer in PDEs numerical simulations with JAX Project: TorchFire - A combination of PyTorch and Firedrake for differentiable machine learning framework Embedded the Firedrake PDE simulations within PyTorch to form an end-to-end differentiable training framework Distributed Firedrake PDE simulations within PyTorch on multiple CPUs for faster training neural networks Project: Data-informed active subspace regularization framework for inverse problems Identified the data-informed active subspace of parameter leading to higher quality solutions compared to conventional frameworks Reduced drastically effort on tuning regularization parameters in Tikhonov framework, leading to a more robust inverse framework Ho Chi Minh City University of Technology July 2017 - Dec 2020 Graduate Research Assistant Ho Chi Minh, Vietnam Implemented direct design method for portal steel frames using structural analysis according to ANSI/AISC 360-16 Developed finite element method programs for non-linear plastic analysis of 2D/3D steel semi-rigid frames Publications H.V. Nguyen, et. al. Learning noisy dynamics of spinodal decomposition from stochastic differential equations. On preparation H.V. Nguyen, et. al. Taen: a model-constrained Tikhonov autoencoder network for forward and inverse problems. Computer Methods in Applied Mechanics and Engineering (2025) H.V. Nguyen, et. al. A model-constrained discontinuous Galerkin network (DGNet) for compressible Euler equations with out-of-distribution generalization. Computer Methods in Applied Mechanics and Engineering (2025) H.V. Nguyen, et. al. Physics-informed dual implicit neural representations for FWI with salt. International Meeting for Applied Geoscience & Energy (2025) H.V. Nguyen, et. al. JAX acceleration of physics-informed FWI and field data application. International Meeting for Applied Geoscience & Energy (2024) H.V. Nguyen, et. al. Tnet: a model-constrained deep learning approach for inverse problems. SIAM Journal of Scientific Computing (2024) R.S. Philley, H.V. Nguyen, et. al. Model-constrained empirical bayesian neural networks for inverse problems. American Congress on Computational Methods in Engineering (2023) J. Wittmer, H.V. Nguyen, et. al. On unifying randomized methods for inverse problems. Inverse Problems (2023) H.V. Nguyen, et. al. A model-constrained tangent slope learning approach for dynamical systems. International Journal of Computational Fluid Dynamics (2022) H.V. Nguyen, et. al. A data-informed active subspace regularization framework for inverse problems. Computation (2022) H.V. Nguyen, et. al. Large Displacement Elastic Analysis of Planar Steel Frames with Flexible Beam-to-Column Connections Under Static Loads by Corotational Beam-Column Element. Journal of Science and Technology in Civil Engineering (2019) H.V. Nguyen, et. al. Large Displacement Elastic Static Analysis of Semi-Rigid Planar Steel Frames by Corotational Euler--Bernoulli Finite Element. Journal of Science and Technology in Civil Engineering (2019) Conferences Invited talk - SIAM Texas-Louisiana Section, TX, USA, 2025 Poster presentation - 5th International Meeting for Applied Geoscience & Energy (IMAGE), USA, 2025 Invited talk + Symposium organizer - 18th U.S. National Congress on Computational Mechanics, USA, 2025 Invited talk - SIAM Conference on Computational Science and Engineering, USA, 2025 Invited talk + Poster presentation + Symposium organizer - SIAM Conference on Mathematics of Data Science, USA, 2024 Invited talk + Poster presentation - SciML Workshop on Scientific Machine Learning, USA, 2024 Invited talk + Symposium organizer - 17th U. S. National Congress on Computational Mechanics, USA, 2023 Invited talk - Seminar at Department of Mathematics - Kansas State University, USA, 2023 Invited talks + Poster presentation + Symposium organizer - SIAM Texas-Louisiana Section, USA, 2022 Invited talk - SIAM Conference on Mathematics of Data Science, USA, 2022 Invited talk - Thematic Conference on Uncertainty Quantification for Machine Learning Integrated Physics Modeling, USA, 2022 Invited talk - SIAM Conference on Uncertainty Quantification, USA, 2022 Teaching Experience The University of Texas at Austin Aug 2021 - Present Teaching assistant Austin, TX, USA Courses: Engineering Computation (COE311K), Software Design and Engineering (COE332), Software Design For Responsible Intelligent Systems (COE379L), Analytical Methods I & II (ASE380P1, ASE380P2), Introduction to Machine Learning (EM397) Delivered tutorials and mentored students about course-related materials including lecture notes, homework, and projects Prepared and evaluated homework problems, exam problems, and project topics Ho Chi Minh City University of Technology Jun 2017 - Nov 2020 Teaching assistant Ho Chi Minh, Vietnam Course: Steel Structures Theory and Design Mentoring Experience William Cole Nockolds, The University of Texas at Austin, Spring Semester 2023 - A model-constrained tangent learning approach for dynamics systems on latent space, pursued PhD in the same group from Sep 2024 Wesley Lao, The University of Texas at Austin, Fall Semester 2022 - Graph neural network model-constrained tangent learning approach for discontinuous wave propagation PDEs using JAX-Fluid package, pursued PhD in the same group from Sep 2023 Nghia Nim, New York University, Summer Semester 2022 - Model-constrained machine learning approach to solving PDEs, Software Engineer at Bloomberg Hieu Tran, DePauw University, Summer and Fall Semesters 2022 - A convolution neural network model-constrained tangent learning approach for dynamics systems, pursued PhD in Purdue University Referees Prof. Tan Bui-Thanh Aerospace Engineering and Engineering Mechanics Department, University of Texas at Austin tanbui@ices.utexas.edu Dr. Anusha Sekar Computation Reservoir Geophysics R&D, Chevron Cooperation anusha.sekar@chevron.com Prof. Cuong Ngo-Huu Civil Engineering Department, Ho Chi Minh City University of Technology cuongngohuu@hcmut.edu Dr. Eng. Takuya Murakami Steel Research Laboratory, JFE Steel Corporation tak-murakami@jfe-steel.co.jp