Match Algorithm

We compare structured resume sections, role context, skills, seniority signals, and job requirements. We also normalize wording so equivalent terms can still match.

The final score combines multiple weighted components: skills-match, role-relevance, experience-depth, and requirement coverage. Each component contributes proportionally to a 0-100 score.

We use vector-based similarity search to compare candidate context and job intent at semantic level, not only exact keywords. This helps detect relevance when wording differs across resumes and descriptions.

If required information is missing, the model lowers confidence and reduces score impact for uncertain components. This prevents overrating sparse profiles.

The match score is decision support, not an automatic rejection system. Recruiters should review context, portfolio quality, and communication fit before making final decisions.