An open R&D challenge across five themes, run on Quantinuum's software stack. This project enters it with an NHS-first framing: the 10 Year Health Plan supplies the problems, the challenge themes supply the instruments.
The themes below are taken from the published challenge material and should be re-checked against the live Aqora page before anything is submitted. What is not provisional is the mapping: each theme is attached to a specific NHS problem and to a gate on this site that already runs.
Circuit-level building blocks: phase estimation, amplitude encoding, rewriting and canonicalisation, tomographic verification.
Verification discipline. If a machine-generated artefact is going to inform care, the harness that proves it correct matters more than the artefact.
The rewriter is proven on 1q and 2q identities. Anything wider is untested.
Electronic structure: ground-state and gap estimation on real molecules, with noise mitigation good enough to trust the number.
Prevention eventually means better medicines. Molecular simulation is the only item on this site with a credible long-run quantum advantage — and the furthest from a patient.
Two qubits. Scaling to a druglike molecule is the entire unsolved problem.
Machine learning in the quantum stack: generative circuit synthesis, learned compilation, RL-refined ansätze.
The same governance question the NHS faces with clinical AI: a model produced an artefact — how do you know it is right? Here the answer is a dense matrix oracle, which is a luxury clinical AI does not have.
No transformer is actually trained here. The composition pattern is demonstrated; the generator is borrowed.
Encoded computation: colour codes, logical rotations, partial fault tolerance, and the mitigation ladder beneath it.
The reliability floor. Any claim about an operational NHS decision needs an error bar that survives scrutiny, and QEC is where quantum computing learns to produce one.
Mitigation, not correction. The Steane-code work on this site is not yet run end-to-end under noise.
QUBO and Ising formulations, variational optimisation, and honest benchmarking against classical solvers.
Discharge flow. Up to one in three beds held by a patient who is medically fit to leave is the NHS's most expensive combinatorial problem — and the one where a solver most obviously is not the missing piece.
Small instances only. The crossover size at which a quantum method would matter is stated, not reached.
The strongest thing here is not a speedup — there isn't one. It is that a clinician-stated NHS problem has been carried all the way to a compiled circuit, run on the vendor emulator, benchmarked against classical solvers that beat it, and shipped with the shot counts and seeds that produced the numbers. That pipeline is reusable for any of the five themes.
The 10 Year Health Plan's three shifts, restated as computational problems.
Read →The flagship: delayed discharge with an equity term, 36 QAOA cells on Selene.
Read →The underlying rewriting and verification programme, gates G1 to G21.
Read →The 22-card Guppy/Selene skill pack, packaged for Lovable, Hermes and OpenClaw.
Read →