UniversalAGI is building foundation AI models for physics that enable end-to-end industrial automation from initial design through optimization, validation, and production. The company is looking to hire a Founding AI + CFD Researcher to build foundation AI models that don't just automate CFD, but fundamentally reimagine how physics simulation works.
Requirements
- 2+ years of hands-on experience building and training deep learning models for scientific computing, physics simulation, or related domains (GNNs, GCNNs, Transformers, Vision Models, Neural Operators, PINNs)
- Strong foundation in CFD: Deep understanding of fluid mechanics, numerical methods, mesh generation, boundary conditions, and solver frameworks
- Proven ML research ability: Track record of implementing novel architectures, running large-scale experiments, and iterating quickly based on results
- Expert-level coding skills in Python and deep learning frameworks (PyTorch, JAX, TensorFlow)
- Experience with CFD software (OpenFOAM, Ansys Fluent, STAR-CCM+, or similar) and the ability to generate, process, and analyze simulation data programmatically
- PhD in Machine Learning, Aerospace, Computational Physics, Applied Math, or related field with focus on physics-informed neural networks, graph neural networks, transformers, geometric convolutional neural networks, neural operators, or scientific ML
- Experience with neural operators (FNO, DeepONet, UNet, Transformers, etc.) or graph neural networks for physical systems
Responsibilities
- Develop novel AI architectures for physics simulation: neural operators, graph neural networks, transformers, diffusion models, surrogate models or whatever works best for learning fluid dynamics
- Design and implement training pipelines that can ingest massive CFD datasets and learn to predict flow fields, optimize meshes, or generate designs with accuracy that matches or exceeds traditional numerical solvers
- Bridge physics and ML deeply: Ensure our models respect physical constraints, conservation laws, and numerical stability, embedding your CFD expertise directly into model architecture and loss functions
- Run large-scale experiments on simulation data, iterate rapidly on model performance, and drive our research roadmap based on what actually works
- Work hands-on with CFD tools (OpenFOAM, Ansys, STAR-CCM+) to generate training data, validate model outputs, and understand where traditional simulation struggles
- Collaborate directly with domain experts and customers in automotive, aerospace, and other industries to understand their workflows, pain points, and validation criteria
- Publish and present breakthrough results, internally and externally, as we push the boundaries of what's possible in AI for physics
Other
- Work Directly with CEO & founding team
- Report to CEO
- 5 Days Onsite
- Strong communicator capable of bridging customers, engineers, and researchers, translating between physics intuition and ML architecture decisions
- Outstanding execution velocity: Ships fast, iterates rapidly, and thrives in ambiguity
