JobMatch Checker
October, 2026
- By Ram Nimbalker
The problem:
Matching a resume to job descriptions is either manual and slow, or black-box AI that spits out a number with no explanation. Recruiters don’t trust it. Candidates don’t learn from it.
What I built:
A scoring system that breaks the match into seven dimensions, explains each score, and tells you what’s missing.
Not just “72% match”, but why 72%, and what would make it 85%.
What it scores:
- Skills coverage (required vs nice-to-have)
- Responsibilities similarity (semantic matching)
- Seniority alignment
- Domain fit
- Education requirements
- Location compatibility
- Outcomes alignment
What makes it different:
- Explainable: Every score has a breakdown and improvement suggestions
- Graceful fallback: Works without OpenAI API (deterministic mode, 55% baseline)
- Mode transparency: Response shows which components used AI vs fallback
- Performance tracked: Every response includes timing breakdown
How I built it:
Python + FastAPI backend. OpenAI GPT-4o-mini for parsing, text-embedding-3-small for semantic similarity. Tailwind + Alpine.js frontend. Docker-ready. Rate-limited API.
Screenshots:
