A rule-based candidate-to-role scoring engine built into WORKSPHR Recruitment. It rates every applicant on a transparent 100-point scale across six dimensions: Role Relevance (30), Skills (25), Experience (20), Education (10), Location/Arrangement (8), Salary Alignment (7). Same candidate plus same job equals same score, every time. No ML black box. No hidden bias. No model drift. Every point is traceable to a rule and a piece of profile data.
Built for recruiters who need to defend every shortlist, every rejection, and every interview slot — to a hiring manager, to HR leadership, or to a labour lawyer.
Deterministic = explainable = auditable. Three properties Filipino HR leaders need to defend hiring decisions, especially under DOLE scrutiny or NLRC review.
Same candidate plus same job equals the same score, every time. Score it today, score it next year, score it after a system migration — identical result. No model drift, no retraining surprises, no “why is the score different now?”.
Reproducibility is non-negotiable when a candidate asks “why was I not shortlisted?”.
Every point on a candidate's 87/100 scoretraces back to a rule and a piece of profile data. “Skills 22/25: found Python, FastAPI, PostgreSQL; missing Docker.” A hiring manager can read the breakdown in 10 seconds and agree or disagree.
If you can’t explain a rejection, the candidate has every right to question it.
Every score is timestamped, with the rule version that produced it, written to score_history. Tune your weights tomorrow — historical scores are frozen for the audit. RA 10173 (DPA) and DOLE inspections handled.
“What were our weights when we rejected this person in March?” Answerable in one query.
Most ATS platforms claim “AI matching” but can't tell you why a candidate scored what they scored. Under RA 10173 (Data Privacy Act)a data subject can demand the logic of any automated decision affecting them. “Our model said so” is not a defence.
Each dimension has a maximum, a clear rule, and a profile field it reads from. Tune the weights, change the score — never change the logic behind it.
A simple, traceable formula. Each candidate scores against the job posting on six dimensions, weighted exactly as shown. The total is a 0–100 integer stored on the job_applications row, returned by every Match API endpoint, and shown on every candidate card.
Match scoring runs automatically the moment a candidate applies. The 100-pt score, breakdown, and insight strings are written to the job_applicationsrow in < 200 ms.
The score is shown on every Kanban tile, candidate card, and pipeline view. Sort by score, filter by score band, bulk-action by score. Score never auto-rejects — recruiters always make the call.
Every score is timestamped with the rule version. Tune your weights tomorrow — historical scores are frozen. RA 10173 (DPA) and DOLE inspections handled with one query.
job_applications row. Every recompute writes a new row to score_historywith timestamp + reason — the old score is never lost.recruitment.match.config:write.job_applications row and shown in every pipeline view, candidate card, and Kanban tile. You can sort, filter, and bulk-action by score. The score never auto-rejects; it only ranks. Auto-reject was a deliberate non-feature — we don't want recruiters or candidates blaming the system.show_salary is true. Segmented by role, location, and experience. Minimum 30 data points before a P25-Median-P75 band is shown. We do not pull from third-party salary databases or scrape job sites. Permission required: recruitment.match.salary_insights.Book a 30-minute demo. We'll score a real-world candidate against your hardest open role, walk the breakdown, and tune the weights live with you.