Helping Barons match CVs to roles with machine learning

Screening CVs can take recruitment teams hours or days. Barons wanted a system that could help rank candidates against job requirements using machine learning.

The technical challenge was to turn unstructured CV text into useful data with a limited supply of labelled training examples, while keeping the application responsive and dependable.

We cleaned the CV data and used natural language processing experiments to extract relevant features. We developed two machine learning models for candidate matching.

The work also covered UX and visual design, frontend and backend development, and a microservices architecture to bring the models into an application.

The delivery combined CV data processing, two matching models and an application interface, giving Barons a system for ranking candidate data against recruitment needs.