Multi-agent CV analysis and enhancement system with CrewAI, DeepSeek, and OpenAI that cuts profile selection time by 70%.
The AI-Powered Resume Optimization system uses a multi-agent CrewAI pipeline to analyze, score, and rewrite CVs for alignment with specific job descriptions — reducing the time HR teams spend on profile selection by 70%.
Three specialized AI agents collaborate: an Analyzer agent extracts skills and experience from CVs, a Scorer agent computes semantic match with the job description using embeddings, and a Rewriter agent uses DeepSeek API and OpenAI GPT to generate an optimized version of the CV.
Built with a Streamlit frontend, the system allows HR teams to upload CVs in bulk, define job requirements, and receive scored, ranked, and rewritten profiles. ChromaDB is used for semantic similarity computation at scale.
Extracting structured text from PDF and DOCX CVs using PyMuPDF and python-docx.
Defining agents (Analyzer, Scorer, Rewriter) with roles, goals, backstory, and tool access.
Embedding CVs and job descriptions with ChromaDB. Computing cosine similarity for match scores.
Prompting DeepSeek/OpenAI to rewrite CVs to better match the job description while preserving truth.
Producing structured reports: match score, skill gap analysis, and optimized CV in PDF format.
Building a multi-page Streamlit app for batch CV upload, JD input, scoring dashboard, and download.