A full-stack generative AI application that analyzes a candidate’s resume against a job description, identifies skill alignment and gaps, and provides structured, recruiter-style feedback. The system is live deployed and usable end-to-end.
Frontend (Vercel): 👉 https://resume-jd-analyzer-jhj7.vercel.app/
Backend API (Render – Swagger Docs): 👉 https://resume-jd-analyzer-backend.onrender.com/docs
This application allows users to upload a resume and a job description in PDF format and receive an AI-driven evaluation of how well the resume matches the role.
The analysis focuses on evidence-based skill matching rather than keyword comparison.
- PDF upload for resume and job description
- Automated text extraction from documents
- AI-based structured resume analysis
- Overall match score (0–100)
- Strong skill matches with evidence
- Missing skills based on job requirements
- Weakly supported skills
- Actionable improvement plan
- Detailed recruiter-style final feedback
- Fully deployed frontend and backend
Frontend:
- React (Vite)
- Deployed on Vercel
Backend:
- FastAPI (Python)
- PDF parsing using pdfplumber
- Async API design
- Deployed on Render
AI:
- OpenRouter (OpenAI-compatible API)
- Model: meta-llama/llama-3-8b-instruct
- Strict JSON-based output validation
React Frontend (Vercel)
→ FastAPI Backend (Render)
→ OpenRouter LLM API
The frontend sends PDF files and metadata to the backend.
The backend extracts text, performs AI analysis, validates the output schema, and returns structured results to the frontend.
POST /analyze
Inputs:
- resume_pdf (PDF)
- jd_pdf (PDF)
- experience_level
- role_type
- industry
- resume_source
Response:
- overall_match_score
- strong_matches
- missing_skills
- weak_evidence
- improvement_plan
- final_summary
Swagger documentation is available at /docs.
- Strict schema validation for AI outputs
- Graceful fallback handling for model errors
- Clean API contracts between frontend and backend
- Environment variables used for secrets (not committed)
- Designed for real-world usage, not just demos
Backend: cd backend pip install -r requirements.txt uvicorn main:app --reload
Frontend: cd frontend npm install npm run dev
Amarnath Devaraj
Built as part of a full-stack AI technical assessment.