GPU Simulation · Founded 2025 · Chicago & Mumbai
Explore more designs.
Wait less for results.
We build GPU-accelerated simulation and optimization software that helps engineering teams run large design-space studies — work that normally sits in a queue for weeks — in a fraction of the time.
Early-stage pilot program open for aerospace, automotive, energy, and electronics engineering teams.
Our 9 Core Engineering Solvers
Unique physics surrogates, finite element analyses, and quantum-inspired combinatorial optimizers arranged in 3x3 layout.
1. QodeX AeroFlow AI
Neural surrogate solver for aerodynamic flow sweeps with sub-second drag and lift prediction.
2. QodeX ThermoCore
Conjugate heat transfer optimization for EV battery modules and liquid cold-plate sizing.
3. QodeX QuantumOptima
Quantum-annealing powered by NVIDIA cuQuantum solving Pareto frontiers across 10,000+ variables.
4. QodeX StressMatrix
GPU-accelerated structural topology optimization powered by NVIDIA Ada Lovelace architecture.
5. QodeX AcousticWave
Broadband NVH and Helmholtz acoustic emission simulator for aerospace propulsion and cabins.
6. QodeX HydroDynamics
Free-surface Volume of Fluid and cavitation solver for ship hulls, foils, and submerged turbines.
7. QodeX NeuroMesh
Deep Fourier neural operator running on NVIDIA H100 GPUs predicting 3D volumetric fields in real-time.
8. QodeX PowerGrid Sim
High-voltage electromagnetic transient solver for inverters, transformers, and grid harmonics.
9. QodeX CloudCluster
Distributed orchestration scheduling sweeps across NVIDIA DGX SuperPOD and Quantum-2 InfiniBand clusters.
How to Use NVIDIA with QodeX Quantum
From local RTX desktop workstations to distributed 512-GPU cloud clusters, here is how engineering teams connect, accelerate, and visualize multi-physics simulations across the NVIDIA hardware and software ecosystem.
Select Compute Target
Run locally on an on-premise NVIDIA RTX 4090 or RTX 6000 Ada workstation for confidential CAD screening, or burst out to cloud NVIDIA H100 SXM5 and B200 clusters with 1-click orchestration.
Compile with TensorRT
Feed raw STEP/IGES CAD geometries. QodeX compiles physics-informed Fourier Neural Operators (FNO) into NVIDIA TensorRT FP16/FP8 engines, delivering sub-second predictions without re-meshing.
Optimize with cuQuantum
Formulate rugged design trade-offs (e.g. drag vs lift vs mass) into QUBO matrices. The NVIDIA cuQuantum state-vector simulator maps true Pareto frontiers across 10,000+ design variables in minutes.
Render in Omniverse
Stream high-resolution volumetric 3D flow velocities, thermal gradients, and von Mises stress contours directly into NVIDIA Omniverse via OpenUSD for real-time digital twin collaboration.
# Step 1: Connect to local RTX or remote NVIDIA H100 cluster import qodex_quantum as qx cluster = qx.DeviceCluster( target="nvidia-h100-sxm5", nodes=4, interconnect="nvlink-800gbps" ) # Step 2: Compile Fourier surrogate with NVIDIA TensorRT FP16 solver = qx.solvers.AeroFlow( precision="tensorrt-fp16", cuda_streams=8 ) results = solver.sweep(geometry="transonic_wing.step", mach=(0.7, 1.4)) # Step 3: Solve Pareto frontier via NVIDIA cuQuantum Annealing optimal = qx.optimizer.cuQuantumAnneal( dataset=results, objectives=["min_drag", "max_lift", "min_mass"] ) # Step 4: Stream 3D fields to NVIDIA Omniverse Digital Twin optimal.export_omniverse(stage="omniverse://nucleus/aerotwin.usd") print("✓ Optimization completed in 1.48s on NVIDIA H100 Cluster")
On-Premise Desktop Workstations
Supports local confidential runs on NVIDIA RTX 4090 and RTX 6000 Ada GPUs with zero cloud upload required.
Cloud GPU SuperPOD Clusters
Elastic multi-node scaling across NVIDIA H100 and B200 nodes connected via 800 Gb/s Quantum-2 InfiniBand.
Native Solver Interoperability
Seamless two-way bridge with Ansys Fluent, OpenFOAM, Nastran, and Abaqus for automated solver verification.
Explore QodeX Quantum
How it works →
The four-step pipeline: define the design space, train fast surrogates, search intelligently, verify and decide.
Use cases →
CFD aerodynamic exploration, thermal and structural trade studies, equipment sizing, and early concept screening.
Technology →
The GPU-first stack underneath: CUDA solvers, physics-informed surrogates, TensorRT inference, cloud GPUs.
Company →
Who we are, where we work, and how to reach the founder and engineering teams directly in Chicago and Mumbai.
Led by World-Class Computational Physicists & GPU Architects
Founded by alumni and researchers from IIT Delhi, IIT Bombay, IIT Kanpur, IIT Madras, and University of Chicago.
Kanak Mehta
14+ years pioneering computational physics software and quantum-inspired optimization engines.
Simran Juneja
Principal GPU architect specializing in CUDA kernel optimization and multi-GPU cluster scaling.
Dr. Ashok Gadgil
National Medal of Technology laureate directing multiphysics modeling, CHT, and solver rigor.
Get in touch with engineering
Interested in evaluating our 9 simulation engines for your design challenges? Join our pilot program.
Contact Engineering Team