About NEROX
Built by optimizers,
for optimizers.
We spent years frustrated by the gap between what quantum-inspired optimization can do in theory and what it actually delivers in production. So we built NEROX.
4,096
Parallel annealing chains per A10G
<2%
Avg gap-to-optimal on TSPLIB-1000
50k
Max variables with Hybrid Solver
5
Solver algorithms shipped at launch
Our mission
Make hard problems
solvable by everyone.
Combinatorial optimization is everywhere — in the routes delivery trucks take, in how hospitals schedule surgeries, in how financial institutions balance portfolios, in how manufacturers sequence jobs on factory floors.
These problems have exact solutions — in theory. In practice, the combinatorial explosion makes brute force impossible. Heuristics work, but leave significant value on the table. Quantum-inspired algorithms close that gap. We exist to make them accessible to every engineering team, not just those with research budgets.
What we believe
Correctness over speed
An answer that's 40% above optimal isn't an answer — it's a liability. We obsess over solution quality, not just throughput. Every solver ships with gap-to-optimal benchmarks against known best solutions.
GPU-first, always
The GPU is the most powerful general-purpose compute primitive available at commodity prices. We build every algorithm to exploit it maximally — parallelism, memory bandwidth, and all. CPU is a fallback, not a target.
No quantum theater
Quantum computing is real and advancing. But most "quantum" software today is marketing. We are honest about what runs on simulators, what requires actual quantum hardware, and what the performance difference actually is.
Accessible interfaces
Operations research should not require an OR PhD. We build APIs, SDKs, and UIs that a software engineer can use without understanding the Ising Hamiltonian. The complexity lives in the solver, not the interface.
The team
A small team with a big problem.
We're engineers and researchers who've spent careers in logistics, finance, and scientific computing — repeatedly frustrated by the state of the art in combinatorial optimization. NEROX is the tool we kept reaching for that didn't exist yet.
Engineering
Research
Product
Infrastructure
We're hiring. If you're obsessed with optimization, GPU programming, or making hard math usable, get in touch.
Ready to optimize?
Start at $19/mo with CPU solvers. Scale to GPU clusters when you're ready.
GPU-native quantum-inspired optimization for combinatorial problems at any scale.