SplitFXM

A platform for physics-aware AI that builds digital twins and accelerates industrial simulations. It solves the problem of complex, high-fidelity simulation by using scientific machine learning and high-order numerical methods. It targets developers, data scientists, and operations teams in engineering and industrial sectors, including mid-market and enterprise companies. It is delivered as a SaaS platform with an infrastructure API, and it differentiates by working directly with governing equations in their natural form, avoiding the need for transformation to first-order systems.

Key features

• Scientific machine learning with DeepONets, FNOs, World Models
• Ultra-high precision with high-order numerical methods
• State-of-the-art optimization solvers for stiff non-linear models
• High-performance parallel execution on CPUs, NVIDIA, AMD GPUs
• Extreme event tracking with WENO/TENO schemes
• Adaptive mesh refinement for dynamic computational grids
• Sparse Jacobians at runtime for large-scale digital twins
• Split methodology for dividing complex systems
• SplitNewton++ core solver for non-linear systems
• SplitContin++ numerical continuation solver
• SplitDAE++ differential-algebraic equation solver
• SplitOPS++ operator-splitting framework