One thing that keeps coming up when talking to ATS builders: candidate matching is basically keyword counting under the hood.
Someone applies → system checks if their resume has the same words as the JD → scores them.
This misses a lot. A senior engineer who wrote "led distributed systems work" isn't going to match a JD that says "distributed systems experience required" — even though it's obviously the same thing.
/v1/score is built for this. You pass a job description + a resume and get back a structured score with breakdown by category: skills fit, experience level, requirements match. It understands the content, not just the tokens.
Built this into JD Intelligence so any platform can plug it in via REST without training their own model or setting up ML infra.