Transparent Match is a job-resume matching system we built for our COMP3901 capstone. Rather than scanning for keywords, it evaluates candidates against job postings across 160 competency dimensions from the O*NET occupational framework — things like reasoning ability, domain knowledge, and how someone handles pressure.
When a resume or job posting is uploaded, a language model assigns competency levels across those dimensions. The system then compares them against job requirements, weighted by how much each competency matters to the role. If the profile has gaps, the system generates clarifying questions. The result is a scored match with a qualification tier and a plain-language explanation — written differently depending on whether a recruiter or candidate is reading it.
On the technical side: FastAPI and PostgreSQL in the backend, React on the frontend with role-based views. We use LiteLLM for the language model calls so we can swap providers without rewriting the extraction logic.
Github Repo: https://github.com/TiannaBassaragh/Capstone-Job-Matching-System
Zip File: https://limewire.com/d/inS7b#Q2Hg6ycjML (Expires on May 29, 2026. Use Github Repo)
Marketing Website: https://transparentmatch.framer.website/