Das Video wird erst beim Abspielen von YouTube geladen.
Mentor Matching Rules and Fit Scores
2:16 Auf Englisch Veröffentlicht am 29. Oktober 2026 Für Programmteams
Add a matching rule live and see how it changes the fit scores. Compare a high-fit and a low-fit mentor–mentee pair, and pick a match from a ranked list.
Transkript
- 0:04 This program has 25 mentors and 75 mentees. Every profile carries real context: role, department, experience, location, and what each person wants from mentoring.
- 0:17 On the matching algorithm page, you decide what makes a good match. Standard rules compare profile fields. Here, what a mentor can help with should be similar to what the mentee wants to work on, and they should come from different departments.
- 0:32 Each rule has a weight, so the things that matter most count the most.
- 0:37 Let's add one more. People in the same city can meet in person, so we pick location for the mentor, and location for the mentee.
- 0:48 The rule is: same. It isn't mandatory, but we give it a weight of two, then add it to the list.
- 0:56 Custom rules handle conditions, like: if the mentor works in engineering, the mentee should be in data. And the AI rule reads what people write in their own words. That one has its own video.
- 1:08 Now we calculate. Every pair is scored with all the rules together, and each mentor's matching limit is respected.
- 1:16 The summary shows who received a proposal. Nothing is final yet. Every result waits for your approval.
- 1:25 Open a proposal to see why it was made. This pair scores high. The green fields are the rules that matched: similar goals, different departments, and the same city.
- 1:36 This one scores lower. The red fields show exactly which rules didn't match, so you decide with the full picture.
- 1:45 You can approve or reject each proposal, or approve them all at once. Depending on your settings, both people get an email.
- 1:54 You can also match by hand. Open a mentor from the matching list, and every mentee is ranked by the same rules. Sort by score, and you see the best fits and the weakest ones.
- 2:10 The algorithm suggests. You stay in control of every match.
