Neil Ashton
Distinguished Engineer, NVIDIA
Biography
I am a Distinguished Engineer at NVIDIA, working at the intersection of computational engineering and artificial intelligence. My research spans computational fluid dynamics, high-performance computing, scientific machine learning, open datasets, foundation models for physical systems, and reliable agentic systems for engineering design and analysis.
I founded the Automotive CFD Prediction Workshop (AutoCFD) and serve on the organising committee of the AIAA CFD High-Lift Prediction Workshop (HLPW), where I lead the AI/ML Technical Focus Group. These programmes coordinate shared test cases, validation methods, and research priorities across academic and industrial groups. I also lead and collaborate on the open AhmedML, WindsorML, DrivAerML, and HiLiftAeroML datasets.
Previously, I was Worldwide Technical Lead for CAE at Amazon Web Services and a Senior Researcher in the Department of Engineering Science at the University of Oxford. Earlier research and engineering work covered industrial CFD, high-performance computing, automotive and aerospace aerodynamics, Formula 1, FIA technical regulation, and British Cycling’s bicycle development programme for the Tokyo 2020 Olympic Games. I completed my PhD at the University of Manchester and am a Fellow of the Institution of Mechanical Engineers and a Chartered Engineer.
Research agenda
My current research is organised around three connected questions.
Foundation models for physical systems
How can models learn from high-fidelity simulations and generalise across geometries, operating conditions, and engineering tasks?
Scientific data and evaluation
Which datasets, benchmarks, and validation methods demonstrate physical fidelity, computational efficiency, and engineering utility?
Agentic engineering
How can AI agents select and orchestrate simulation, analysis, and optimisation tools, verify physical consistency, preserve provenance, respond to uncertainty and failures, and keep consequential decisions under appropriate human oversight?
Current research
Research overviewCurrent work combines open scientific data, high-fidelity simulation, foundation models, and agentic engineering.
Open scientific data
CAE ML Datasets
I lead and collaborate on open, high-fidelity datasets—including AhmedML, WindsorML, DrivAerML, and HiLiftAeroML—for reproducible automotive and aerospace research.
Project websiteFoundation models
Fluid Intelligence
A research perspective and scaling analysis for foundation models in computational fluid dynamics, including the respective costs of data generation and model training.
PaperAgentic engineering
Agentic computational engineering
I research agentic AI systems that formulate engineering problems; select and orchestrate simulation, analysis, and optimisation tools; verify physical consistency; preserve traceable provenance; respond to uncertainty and failures; and support human engineering judgement.
NASA Ames talkCommunity workshops
I founded AutoCFD and serve on the HLPW organising committee, where I lead the AI/ML Technical Focus Group. These programmes establish shared test cases, evaluation methods, and community research priorities for computational aerodynamics.
Automotive CFD Prediction Workshop (AutoCFD)
Founder and organiser
An international workshop series coordinating community assessment and validation of CFD methods for road-vehicle aerodynamics.
AIAA CFD High-Lift Prediction Workshop (HLPW)
Organising committee and AI/ML Technical Focus Group lead
A collaborative workshop series advancing computational methods and validation practices for complex high-lift aircraft configurations.
Selected publications
All publications2026
HiLiftAeroML: High-Fidelity Computational Fluid Dynamics Dataset for High-Lift Aircraft Aerodynamics
arXiv preprint
An open high-fidelity dataset comprising 1,800 samples across 180 geometry variants and ten angles of attack.
2025
Fluid Intelligence: A Forward Look on AI Foundation Models in Computational Fluid Dynamics
arXiv preprint
Develops a CFD-specific scaling law and estimates the data-generation and training costs of foundation models.
2024
WindsorML: High-Fidelity Computational Fluid Dynamics Dataset for Automotive Aerodynamics
NeurIPS 2024, Datasets and Benchmarks Track
A peer-reviewed open dataset of 355 GPU-native wall-modelled large-eddy simulations for automotive aerodynamics.
2024
Summary of the 4th High-Lift Prediction Workshop Hybrid RANS/LES Technology Focus Group
Journal of Aircraft 61(1), 86–115
A multi-team assessment of hybrid RANS/LES methods for complex high-lift aircraft configurations.
2023
Overview and Summary of the First Automotive CFD Prediction Workshop: DrivAer Model
SAE International Journal of Commercial Vehicles 16(1), 61–85
The founding workshop’s assessment of 53 CFD datasets from nine groups using open DrivAer test cases.
2016
Assessment of RANS and DES methods for realistic automotive models
Computers & Fluids 128, 1–15
A systematic comparison of turbulence-modelling approaches for the Ahmed and DrivAer automotive configurations.
Selected invited talks
Talks archive2026
Sponsored keynote
SIGGRAPH 2026
NVIDIA | Next Era of Graphics — Neural Rendering, World Models, and Simulation
Event page2026
Invited seminar
NASA Ames Seminar Series
Perspectives on Foundation Models and Agentic Architectures for Computational Fluid Dynamics
Watch seminar2026
Invited seminar
University of Oxford Mathematical Institute
Towards a Foundation Model for Computational Engineering: Opportunities, Challenges, and Novel Scaling Laws
Event page2025
Invited talk
NASA High-Fidelity CFD Workshop
Enabling industrially relevant high-fidelity CFD and AI surrogate models for external aerodynamics
2023
Invited keynote
22nd International Computational Fluids Conference
The role of cloud computing, machine learning, and HPC in the advancement of high-fidelity CFD for industry
Research updates
-
The HiLiftAeroML preprint is now available, describing the open-source high-fidelity CFD dataset for high-lift aircraft aerodynamics.
-
My recent NASA Ames Seminar is now available online. It covers the Fluid Intelligence paper and perspectives on recent advances in agentic AI. Watch the seminar.
-
I’m excited to release the preprint Fluid Intelligence: A Forward Look on AI Foundation Models in Computational Fluid Dynamics with Johannes Brandstetter and Siddhartha Mishra, exploring the path towards foundation models for CFD.
Contact
For research correspondence, academic talks, or questions about datasets and publications, please contact me by email.