Skillmatch Pro

Publié le 31 octobre 2024

Voir la réutilisation

Effective invention team

Our team took part in the OpenData x Skill hackathon and developed a concept aimed at assisting candidates in showcasing skills that may be obscured within job titles or experience descriptions.

1 réutilisations

  • adem
  • competences
  • demandeur-demploi
  • emploi
  • employment
  • resume
  • skills

Informations

Type
Application
Thématique
Population et société
ID
6723bfdb3e762b1f57870150

Publication

Intégrer sur votre site

URL stable

Description

A tool helping you showcase your most relevant skills on your resume

We want to help users create resume that will attract more attention from HR people

  • Offering to describe their work experience including skills they mobilized
  • Using skill names that are seen in job postings globally
  • Using skill names that are relevant in the market

But also be better and quicker to create tailor made resume for a specific job

  • Translating automatically your resume
  • Rewrite work experience based on the skills you picked up
  • Adapt tone of the resume to a specific sector or job ad

Challenges & Limitations

  • "Stupid AI" (the temperature was way to high at first)
  • Being hired is a “beauty contest” that required to market themself while data from ADEM and ESCO are more factual about the jobs and the skill needed
  • How do you actually prioritize skills? Is there one single way of doing it? Or a best way to do it?
  • Same questions with “What is the best way to label a skill?”
  • Who should decide? AI, market data, or user?
  • Skewed market data
  • Questionable data on both ESCO and market side (e.g. “accounting” highly required by the market for accountants, not related in ESCO)

What could be next?

  • Skill matching experience: display the missing skills one by one and let user swipe / select which one they wants to include or not
  • Offer to generate a PDF with the refinements in place using a standard layout
    • rewrite the experiences descriptions based on the selected skills
    • rewrite in different language
    • better accessible PDF structure (great for automated HR software & humans)
  • Offer the tool in more languages and markets (we focussed on English & Luxembourg market data)
  • Use ESCO Data into a vector storage and ask the LLM to use it to output directly information following ESCO nomenclature

Checkout the slide deck here

Jeux de données utilisés 1

Discussion entre l'organisation et la communauté à propos de cette réutilisation.

Plus de réutilisations

Découvrez davantage de réutilisations.