Fair, stable matching of trainees to clinical institutions

Any medical college can pair its trainees with training posts using the Roth–Peranson deferred-acceptance algorithm — the same family of method behind national medical residency matches. The result is stable: no trainee and institution would both prefer each other over the places they were given. Because the match is run by an independent third party with no stake in the outcome, colleges can present results with confidence that placements reflect stated preferences alone — not human judgement.

Launch the matching tool Request access

Access is by invitation — see below to sign up.

How it works

1

Collect preferences

Trainees rank the institutions they'd like to train at; institutions rank the trainees they'd like to take, and state how many posts they have.

2

Run the match

The algorithm processes both ranked lists together and finds an assignment that respects every institution's capacity and everyone's stated preferences.

3

Get an independently verified result

Every result is automatically checked for stability. Because the algorithm is deterministic, the same inputs always produce the same outcome — and no one can alter a placement without the verification failing.

No conflict of interest

Your college collects trainee and institution preferences — but the matching itself is run by a neutral third party. The Roth–Peranson algorithm is fully automated and has no discretion: given the same ranked lists, it always produces the same result, and the output is mathematically verified to be stable before it is returned to you.

This separation of responsibility mirrors the principle behind national medical residency programmes. Neither the college nor this service can influence individual placements — the algorithm decides, and the stability check proves it.

Why stable matching?

A matching is stable when there is no trainee–institution pair who would both rather be matched to each other than to what they got. Stability is the accepted standard of fairness for this kind of allocation: it removes the incentive to game the system and means no one is left with a justified grievance.

Our engine uses candidate-proposing deferred acceptance, which produces the outcome that is best for the trainees among all stable matchings.

Your data stays private

  • Uploaded files are processed in memory only — input CSVs are never stored.
  • Standard mode (default): nothing is retained after the match completes.
  • History mode (opt-in): match results and the algorithm trace are saved for your organisation's future reference. Input files are still not stored.
  • The service runs in a single, access-controlled region.
  • Identifiers are treated as text, so leading zeros are never lost.

Request access

The matching tool is available to approved users. To set up an account, get in touch and we'll provision a login for you.

Already have access? Launch the tool →