Torre’s Job-Matching Model | Recentness of similar experiences
    • English (en) uk flag arrow_right

    The global remote work community tends to interact in English. Switching languages might impact your visibility.

  • Search
  • Jobs/gigs
  • Publish a job
  • Your jobs (posted or applied)
  • Preferences
  • Your genome
  • Signals
  • Messages
  • Torre's product roadmap
  • Request features
  • API for developers
  • Help
4.9. Recentness of similar experiences
Recentness of similar experiences
4.9.1.
Purpose
The purpose of the Experience Recentness ranker is to compare the candidate's previous job experiences and evaluate how recent and similar they are compared to the job the candidate is applying for.
4.9.2.
Data
For this factor, all of the candidate's previous experiences are considered, excluding those related to education. Each one of them is compared with the title of the opportunity the candidate is applying for, using text similarity algorithms and the recentness of the experience.
4.9.3.
Data validation
Candidate experiences can be verified by other profiles, enhancing their credibility.
4.9.4.
Algorithm validation
  • Daily and weekly metrics
  • Editorial reviews
  • Direct feedback from talent seekers
  • Direct feedback from candidates
4.9.5.
Description
4.9.5.1.
Iterate over each experience of the candidate:
Start by going through each experience except educational ones.
4.9.5.2.
Compute text similarity between role titles:
Compare the current experience's title and the job opportunity's title with levenshtein algorithm and compute a text similarity score.
4.9.5.3.
Filter experiences based on similarity score:
Filter experiences that have a similarity score bigger than the threshold of 40%.
4.9.5.4.
Compute normalized experience recentness:
Retrieve current experience's end date and count the months passed until the job opening's published date. Normalize that amount over 2 years expressed in months.
4.9.5.5.
Retrieve best scoring experience:
Return the experience with the the highest score between similarity and recentness properties.
Fair artificial intelligence.
We’re committed to it.
Fair artificial intelligence.