misterioul/Jobs
🛠 Extract skills from job descriptions and match candidates with relevant resumes using this NLP-powered recommendation engine.
What's novel
🛠 Extract skills from job descriptions and match candidates with relevant resumes using this NLP-powered recommendation engine.
Code Analysis
13 files read · 3 roundsAn NLP pipeline that scrapes job descriptions from Dice.com, uses dictionary-based BIO labeling to create training data, fine-tunes a BERT token classification model for skill extraction, and serves predictions via a FastAPI endpoint with extensive post-processing heuristics.
Strengths
The inference post-processing pipeline (subword merging, 5-case fragment detection, multi-tier confidence thresholds, dictionary complement) shows genuine engineering thought. The end-to-end flow from scraping through training to API is complete and functional.
Weaknesses
No tests exist anywhere in the project. The README is largely fabricated — it describes a downloadable desktop application with Docker/CI-CD/AWS deployment that doesn't exist in the code. Auth uses hardcoded credentials and a hardcoded secret key.
Score Breakdown
Signal breakdown
Innovation
Craft
Traction
Scope
Evidence
Commits
13
Contributors
2
Files
31
Active weeks
6
Repository
Language
Python
Stars
3
Forks
2
License
—