Overview
This project explores how natural language processing can support job seekers without replacing their own judgment. The application scores alignment between a resume and a job description, then shows missing keywords, overlapping strengths, and concise improvement suggestions.
Approach
The first version uses TF-IDF features, cosine similarity, and a curated skill dictionary. A future version can add embeddings and role-specific weighting.
What I learned
The most important design decision was making the explanation more useful than the score. Recruiters and candidates need clear signals, not mysterious percentages.