matching engine or Artificial Intelligence (AI) at the service of recruiters.

Artificial intelligence (AI) is according to the Le Larousse dictionary a set of theories and techniques implemented from complex computer programs, with a view to producing machines capable of simulating certain traits of human intelligence, in particular its cognitive functions.

Machine learning

In this universe, is from its launch oriented towards the logic of rules engine and machine learning. Interviews with recruiters with at least more than 20 years of experience allowed the collection of observations. From these observations, an analysis model was developed. The data taken into account are both professional skills and interpersonal skills, areas of experience, initial and continuing training, geographic location, remuneration, etc. In addition, the team wanted to take all the necessary precautions to prevent any form of discrimination and opted for compliance with the law on equal opportunities (law n ° 2006-396 of March 31, 2006) thanks to the anonymity of applications until the interview. This allows all applicants to be judged and selected solely on objective criteria.

Supervised learning and Q-Learning

This first version is based on so-called supervised learning, with labeled data and discriminant analyzes. However, the increase in the volume of data, in particular the different assessment systems and the choice of recruitment, ultimately orient the model towards a logic of reinforcement learning and Q Learning, which will lead the tool to learn the “Habits” of the recruiter and compare them to the “habits” of other recruiters in the application. This opportunity is made possible by the initial choice to base the data model on a graph (NEO4J) which allows the appearance of unlikely data relationships and helps in the discovery of new model opportunities.

Backlog in effervescence and fundraising

The design roadmap has many avenues to enrich the initial model. The R&D pole is in turmoil and almost every day feeds the project backlog with new ideas and improvements. will therefore soon launch a fundraiser to accelerate its development. The team should thus be reinforced with the skills of Machine Learning-oriented Data Scientists and Python developers.

The startup and its showrooms is a startup publisher of an Applicant Tracking System (ATS) solution, the objective of which is to provide recruiters exposed to permanent application files, neutral and factual assistance in the delicate candidate selection process. In addition, has just opened a first showroom, the ITSkills platform, dedicated to recruitment in the IT world. It is planned to duplicate this model in other sectors of activity at a later date.