Mentor matching,already solved.
PairScore's compatibility framework also supports mentor and mentee matching. The technology is currently being applied in partnership with To Be An M.D. to create context-aware matches while accounting for participant needs, preferences and program capacity.
Four passes.In this order.
- 01
Themes are extracted first
Mentor and mentee narratives are read for experience themes and growth opportunities. What counts is the substance of what someone did, not the category they filed it under. A mentee who wants help getting into research and a mentor who found their way in through a lab are a match on a theme, not on a dropdown.
- 02
Mutual requests take priority
If a mentor and a mentee have asked for each other, that pairing is protected. Your coordinators usually know something the algorithm does not, and an optimization that quietly breaks a relationship somebody already built is not an improvement.
- 03
Stable matching runs over the whole pool
Not first come, first served, and not one pairing at a time. Stable matching produces an assignment where no mentor and mentee would both rather have been paired with each other than with who they got. That property is what stops the last few people in an alphabetical list from getting whatever is left.
- 04
An AI shortlist pass grades the finalists
For each mentee, the strongest candidates are evaluated against one another, and the AI Alignment figure reports how good the chosen pairing is relative to the other combinations the pool could have produced. It is a quality read on the match, not a personality assessment of the people in it.
What a run returns.Including who missed.
An unmatched participant is a result, not an error, and the run tells you which constraint produced it: no eligible mentors, no remaining capacity, or a pairing blocked by your own role rules. That list is usually the most actionable thing in the whole report.
Match Run · Sample cohort
Completed
- Total pairs
- 34
- Capacity Utilization
- 82%
- Avg match score
- 78
- Unmatched
- 3
- MEN-118MTE-042matched to each other. Mutual request.Mutual request92
- MEN-104MTE-017matched to each other. Shared theme: rural health.Shared theme: rural health88
- MEN-131MTE-055matched to each other. Shared theme: translational research.Shared theme: translational research84
- MEN-127MTE-009matched to each other. Capacity 2 of 3 used.Capacity 2 of 3 used71
Illustrative example. Sample run, not a real cohort.
Three numbersyou will be asked.
- PairScore
- Your overall compatibility score for a mentor-mentee match based on factors such as shared goals, interests, communication preferences, experiences, mentorship needs, and overall alignment patterns. The score is designed to provide a more holistic view of mentorship fit beyond just surface-level similarities.
- Capacity Utilization
- Shows how much of the available mentor capacity is currently being used within a match run or cohort.
- AI Alignment
- Uses AI to evaluate the quality of a mentor-mentee match relative to other possible pairing combinations within the participant pool.
Tell us how you match.See the difference.
Send us your cohort size and how pairing works in your program now, and we will walk you through what a Match Run would return.