🧠 Resume Matcher AI

A Streamlit web app for recruiters and hiring teams to evaluate student resumes using AI-powered matching logic. This tool helps identify top candidates based on hard skills, contextual experience, and soft skill inference — not just keyword stuffing.

---

🚀 How It Works

Upload a batch of student resumes (PDF/DOCX) and paste in a job description. The app uses OpenAI's GPT model and custom logic to rank each resume against three evaluation levels:

---

🎯 Match Model: 3 Levels

Level	Description	Weight
Level 1: Skills Match	Does the resume include the required hard skills or tools mentioned in the job description? Uses deterministic keyword matching for accuracy.	60%
Level 2: Experience Relevance	Does the candidate have project or internship experience that aligns with the goals and intent of the job (e.g., building tools, APIs, systems)? GPT is used to interpret project relevance, not just tool names.	25%
Level 3: Soft Skill Inference	Does the resume demonstrate communication, clarity, initiative, or project leadership? GPT evaluates how well projects are articulated (STAR method, metrics, portfolio links, etc.) and how the resume is structured.	15%

Each level outputs:

• A score (0 to 1)
• A bullet-point rationale
• A combined weighted score used to rank all candidates

---

📁 Input Format

• Upload resumes as .pdf or .docx
• Paste a job description in the text box
• App will return:
  • Match Score
  • Level 1, 2, and 3 Scores + Rationale
  • Extracted email
  • Profile URL (LinkedIn preferred, GitHub as fallback)

---

📤 CSV Export

Each run generates a downloadable .csv containing:

• Resume filename
• Match Score
• Level 1 / 2 / 3 scores
• Rationale (as bullet points)
• Extracted email
• Extracted profile URL (LinkedIn > GitHub)

---

🛠 Local Setup

pip install -r requirements.txt
streamlit run app.py