Eliminating the Simulation Bottleneck
Quantum Bridge AI AB is a Deep Tech engineering firm dedicated to bridging high-fidelity academic physics and real-time industrial optimisation. We replace empirical trial-and-error with mathematical certainty.
The Mission
We build proprietary computational physics engines that allow companies to command their physical processes with mathematical certainty. Our Laminar SciML framework transitions R&D pipelines from slow, legacy solvers to high-speed Neural Surrogates — achieving 100x to 1000x speedups without sacrificing physical fidelity.
Technical Sovereignty
All core solver logic is verified Background IP, formally registered under KTH Innovation Legal (Ticket #KTH-INC-5175285, February 2026). Our ROF Hessian methodology was subjected to rigorous peer scrutiny during an invited technical seminar at the University of Jyväskylä.
From research to enterprise
Foundation Research
JAX-NQS Variational Monte Carlo engine achieves 0.04% error on 1D TFIM. Neural Quantum States validated.
Veritas-JAX Engine
KenCarp4 IMEX solver reaches 10⁻¹⁴ precision. OPC-UA Bridge + TenSEAL CKKS vault integrated. TRL 7 achieved.
Laminar-GND Architecture
O(1) memory continuous Graph Neural Diffusion engine built on JAX/XLA. Benchmarked on Cora (71.3% — metastable equilibrium).
Quantum Bridge AI AB
Company registered in Sweden. Corporate whitepaper published. Background IP registered via KTH Innovation Legal.
Shek Lun Leung
Shek Lun Leung
Computational Physicist & Sole Architect
Specialising in differentiable physics engines, Neural Quantum States, and continuous graph neural architectures. All architectural decisions, solver implementations, and neural configurations originate exclusively from the named developer.
Interested in collaborating?
We welcome partnerships with enterprises, research institutions, and deep-tech investors.
Get in Touch