Our portfolio

Have a look at our works

Fast-text algorithm implementation

The goal

To build a predictive model to match job posts to the job seekers, using text from the job seekers resume and job announcement.

What we did

We implemented Facebook’s fast text library for the task. The model was trained on over 2 mln job posts. The result is a ready-to-go model that calculates the matching score between a job post and a candidate. The following statistical methods were used: Word2Vec, Sentence2Vec, supervised learning.

  • Client: Online job-matching company, USA