Vikramaditya Roy
Pioneering Continuous Spectral Learning for Complex Aerospace PDEs
Vikramaditya Roy is an aerospace systems scientist and entrepreneur with deep expertise in computational fluid dynamics, non-equilibrium aerothermodynamics, and high-performance neural computing. Prior to founding StrataPhysics Inc., he served as lead aerothermodynamics architect developing spectral boundary layer methods and oblique shockwave solvers for high-Mach atmospheric flight.
At StrataPhysics, Dr. Roy leads the development of infinite-dimensional Fourier Neural Operators that solve the multi-billion dollar supercomputing simulation bottleneck. Under his direction, the company has authored custom warp-synchronous GPU tensor routines that solve 33 million continuous spatial points in under 16.8 milliseconds.
Senior Research & Supercomputing Team
Palo Alto researchers in applied mathematics, spectral solvers, and distributed tensor clusters.
Dr. Anya Kasparova
Doctorate in Applied Mathematics. Specializes in hyperbolic conservation laws, Riemann invariants, and physics-informed continuous deep learning.
Tariq Lin
Supercomputing performance specialist. Expert in parallel 3D Fast Fourier Transforms, multi-node optical ring AllReduce algorithms, and FP8 kernel tuning.
Dr. Marcus Lindqvist
Former aerospace institute researcher specializing in Large Eddy Simulation (LES) neural surrogates and boundary layer turbulence modeling.
Provable Mathematical Governance Principles
All neural operator forward passes are constrained via divergence-free projection layers, enforcing continuity.
Proprietary aerospace CAD geometries and classified wind tunnel telemetries are processed without external data leakage.
Every model checkpoint publishes formal spectral decay bounds guaranteeing stability up to Mach 10.