How Seeker Analyzes 168,000+ Job Listings
Every number on this site comes from a real, continuously updated job corpus. Here is exactly how we collect, process, and match jobs to resumes.
Data sources
Seeker ingests job listings from 18 verified sources across three tiers:
Tier 1: Company Career Pages
Trust: 95%Direct API connections to applicant tracking systems. Highest data quality: full descriptions, structured skills, salary data.
Greenhouse, Workday, Ashby, SmartRecruiters, Apple, Amazon, Kaiser Permanente
Tier 2: Curated Job Boards
Trust: 70-80%Aggregators and specialty boards with editorial curation. Good data quality with moderate skill extraction rates.
Himalayas, Adzuna, USAJobs, Jooble, Reed, TheMuse, FindWork, Remotive
Tier 3: Supplemental Sources
Trust: 40-60%Additional coverage for niche domains. Lower trust score, used to fill geographic and domain gaps.
Jobicy, Careerjet, JSearch
Each source has a trust score (0.0-1.0) that reflects data quality, description depth, and skill extraction reliability. Trust scores influence match confidence but do not filter results.
Update frequency
The corpus is updated continuously via automated ingestion cycles:
- Tier 1 sources: Every ingestion cycle (approximately hourly). Change detection skips unchanged boards to reduce API load.
- Tier 2 sources: Every cycle, budget-permitting. API call budgets prevent over-fetching from rate-limited sources.
- Deactivation: Jobs not seen in 14+ days are automatically deactivated. The corpus reflects currently active listings, not historical postings.
Skill extraction and matching
When you upload a resume, Seeker runs a multi-stage analysis pipeline:
- Resume parsing: Extracts structured data from PDF/DOCX including work history, education, skills, and project descriptions.
- Skill extraction: Identifies 15-30 skills per resume using a domain-aware taxonomy. Skills are classified as primary (frequently mentioned, high evidence) or secondary.
- Corpus matching: Compares your skill profile against all active listings simultaneously. Each match is scored across skill overlap, seniority alignment, domain relevance, and title intent.
- Tier assignment: Matches are bucketed into High Priority, Strong Fit, Good Bets, and Trajectory based on composite score and evidence strength.
- Quality gating: Thin listings (fewer than 3 extracted skills or under 200 characters of description) cannot appear in the top tier regardless of score.
Match scoring
Each job match receives a composite score (0-100) built from independent signals:
Skill alignment
40%Direct and transferable skill overlap between resume and job listing.
Seniority fit
25%Level alignment between candidate experience and role expectations.
Title intent
15%Semantic similarity between candidate role history and job title.
Domain relevance
10%Domain distance between candidate background and job category.
Location + modifiers
10%Geographic alignment, remote compatibility, and company affinity.
Scores are not curved or normalized. A 78% match means 78% of the scoring criteria are met by evidence on your resume. Two candidates can both score 78% on the same job through different skill combinations.
What we do not do
- We do not fabricate statistics. Every number on this site is derived from the live corpus or clearly attributed to industry sources.
- We do not sell resume data. Your resume is processed for matching and is not shared with employers, recruiters, or third parties.
- We do not inflate match scores. If your resume does not match a job well, the score reflects that honestly.
- We do not use AI to generate job listings. Every listing in the corpus is sourced from a real employer career page or verified job board.
Corpus snapshot
168,000+
Active listings
18
Verified sources
~49k
New per week
Questions about our methodology? Read our guides or reach out at support@seekerscore.com.