Expertise
Scientific AI and reduced models
Physics-informed networks, neural operators, POD reduced-order models and multi-agent simulation workflows for materials and process design.
Named methods
- physics-informed neural networks
- DeepONet
- Fourier neural operators
- physics-informed neural operators
- proper orthogonal decomposition ROMs
- scientific foundation models
- multi-agent simulation frameworks
- reinforcement learning from verifiable rewards
- retrieval-augmented generation
- knowledge-graph scientific workflows
Contribution to a work package
Flowphys can lead or support work packages that build scientific foundation models, surrogate stages in multi-physics chains, agentic simulation workflows and APIs that expose models to partners.