Challenge

Successful recruitment consultancies rely on collective experience and knowledge of the job market to expertly match job candidates to roles based on relevant skills and experience at similar companies as well as practical qualities like geographical location. Our client asked us to help them build a system capable of identifying the best and most relevant candidates for their customers who submit job specifications by email.

Solution

The client provided us with a number of expert profiles (digitised CVs compiled by hand and from third party sources like linkedin) and they also sent a number of requests for consultants/experts.

We ran an initial data study to investigate whether some common NLP techniques could be used to extract key information such as names of people and companies, relevant skills and relevant industries from job specification emails.

We worked with the client to build a graph database that connects experts to companies, educational institutions, skills and industries. We provided a flexible framework for mapping expert profiles from a range of import sources into the knowledge graph.

We then utilised a number of Natural Language Processing techniques including Named Entity Recognition, Text Classification and Topic Modelling in order to extract key information from the job requests received from recruitment clients. This information was used to generate knowledge-graph queries and retrieve relevant users from the database.

The initial data study to investigate the was carried out over 4 weeks and the full project was implemented over 4 months.