Recruitment AI Matcher

Most recruitment search works on keywords. Type "React developer 3 years" and you get every CV containing those words, in no useful order. An experienced recruiter can spot the right person by reading a CV, but nobody can read 500 of them for one role.

Golden Sea built a platform that does a first read the way such a recruiter would. Every CV and job description is turned into vectors with a tuned embedding model and stored in Chroma. When a new job comes in, the system pulls the CVs closest in meaning, so it finds people whose experience fits the role even when they use different words.

Each candidate is then scored on 12 criteria: technical skills, years of relevant experience, industry, education, languages, location, seniority, leadership, specialisation, size of past projects, signals in their writing, and whether their career is moving up. The scores add up to a match percentage.

Candidates land in three groups: strong match, worth a second look, and not a fit. Recruiters see why each person ranked where they did instead of a bare score. The platform plugs into the agency's existing CV database and LinkedIn sourcing.