A paired kidney exchange is a graph problem where the objective is measured in transplants. Image segmentation guides diagnosis. Classification separates responders from non-responders. All three have quantum formulations you can run in a browser today.
Incompatible donor–recipient pairs form cycles; packing the most transplants into feasible cycles is NP-hard. Our demonstration finds the 7-of-8-transplant optimum on a real exchange topology.
Edges are where anatomy changes. Quantum edge detection encodes an image in amplitudes and finds every edge with a single Hadamard gate — an honest, verifiable primitive.
Some cohorts are provably not linearly separable. A quantum kernel maps them to a space where they are — and we show exactly the case where that gap is provable.
We take one allocation or classification problem from your clinical or operational side and write down its graph, its constraints, and what 'better' means clinically.
The formulation becomes a program in our lab — every optimisation step a real circuit run, every result reproducible by your own team, signed out, free.
A written assessment of scale: what today's machines handle, where classical solvers still win, and what changes as hardware grows — pursue, park, or drop.
These programs are published in our algorithm library. The numbers below come from recorded executions we can reproduce on demand.
Packs transplant cycles in a paired-exchange graph with QAOA.
Measured: Found the 7-of-8-transplant optimum at 5.9× the random-guessing rate.
Finds all edges of an amplitude-encoded image with one Hadamard.
Measured: Output correlated 0.999 with the exact classical gradient.
Classifies with a quantum kernel where linear methods provably fail.
Measured: Separates a dataset no linear classifier can, by construction.
Results are from the library items' own recorded runs on our simulator — the same one your browser uses.
Clinical deployment has a bar that demos do not clear. Our healthcare programs are decision-support mathematics at demonstration scale — they are not medical devices, and we won't blur that line.
Exchange design, scheduling, cohort classification — describe it and an engineer replies within a day.