📄 AI Resume Analyzer

A full-stack system that evaluates a candidate's resume against a job description using semantic search and generative AI.

---

🧠 Overview

This application extracts text from uploaded resumes (PDF), breaks it into manageable chunks, embeds the chunks using a powerful embedding model, stores them in a vector database, and later performs a semantic search when a job description is provided. The retrieved chunks are passed to a generative AI model to create a well-formatted evaluation.

---

🧰 Tech Stack

Component	Technology Used
Embedding Model	text-embedding-004 (Gemini)
AI Model	gemini-flash-2.0
Vector Database	Pinecone
Frontend	React
Backend	Express.js, Node.js
PDF Parsing	pdf-parse

---

🔄 Workflow

1. 📥 PDF Upload & Parsing

• A user uploads a resume PDF.
• The backend uses pdf-parse to extract raw text.

2. ✂️ Chunking

• The extracted text is split into chunks of ~500 words each for better embedding and semantic resolution.

3. 🧬 Embedding

• Each chunk is embedded using Gemini's text-embedding-004 model.
• Embeddings are paired with the original text and stored in Pinecone under a defined namespace.

4. 🔍 Semantic Querying

• When a user submits a job description, it's embedded the same way.
• A semantic similarity search is performed on Pinecone using the embedded query.
• Top-k relevant chunks from the resume are retrieved.

5. 🤖 AI-Powered Evaluation

• The top chunks + job description are fed into gemini-flash-2.0.
• A formatted, markdown-compliant evaluation is generated, analyzing:
  • Strengths
  • Gaps
  • Fit Verdict
  • Recommendations

6. 🌐 Frontend Display

• The final response is returned to the frontend.
• The React app renders the markdown evaluation cleanly for the user.

---

🧪 Example Use Case

1. Upload resume PDF →
2. Query with job description →
3. View detailed HR-style candidate assessment

---