Resume Screening System

About the workflow

# AI-Automation-Workflow-Resume-Screening-System An intelligent, multi-agent workflow automation system that screens resumes using Large Language Models (LLMs), providing explainable rankings and automated candidate communication. ## Key Features - Multi-Agent Architecture: 4 specialized AI agents (Parser, Skill Matcher, Ranker, Email Generator) - Explainable AI: Confidence intervals and detailed reasoning for all rankings - Automated Pipeline: End-to-end processing in 30-40 seconds - Database Persistence: Full audit trail with status tracking - Personalized Communication: Context-aware candidate emails - REST API: Easy integration with existing HR systems ## System Architecture ``` ┌─────────────┐ ┌──────────────┐ ┌─────────────┐ │ Webhook │────▶│ PDF Parser │────▶│ Database │ │ (REST API) │ │ (LLM Agent) │ │ (Postgres) │ └─────────────┘ └──────────────┘ └─────────────┘ │ ▼ ┌─────────────┐ │ Get Active │ │ Job │ └─────────────┘ │ ▼ ┌──────────────────────────────────┐ │ Skill Matcher Agent │ │ (Technical Recruiter AI) │ │ - Match % calculation │ │ - Skill gap analysis │ └──────────────────────────────────┘ │ ▼ ┌──────────────────────────────────┐ │ Ranking Agent │ │ (HR Analyst AI) │ │ - Confidence intervals │ │ - Strengths/Improvements │ │ - Recommendation │ └──────────────────────────────────┘ │ ▼ ┌──────────────────────────────────┐ │ Email Generator Agent │ │ (HR Communications AI) │ │ - Personalized emails │ │ - Context-aware messaging │ └──────────────────────────────────┘ │ ▼ ┌─────────┐ │ Gmail │ │ SMTP │ └─────────┘ ``` ## Technology Stack | Component | Technology | |-----------|------------| | Workflow Automation | n8n | | AI Framework | LangChain | | Large Language Model | Groq (Llama 3.3-70B) | | Database | PostgreSQL (Supabase) | | Email Service | Gmail SMTP | | API | REST (Webhook) | ## Agent Design ### 1. Resume Parser Agent - Role: Extract structured data from unstructured text - Input: Raw PDF text - Output: JSON with 8 fields (name, email, skills, experience, etc.) - Accuracy: 95%+ ### 2. Skill Matcher Agent - Role: Compare candidate skills against job requirements - Input: Candidate profile + Job description - Output: Match percentage, skill gaps - Precision: 90%+ ### 3. Ranking Agent - Role: Provide explainable rankings with confidence - Input: Match score + Full candidate profile - Output: Overall score (0-100), confidence interval, top 3 strengths/improvements, recommendation - Explainability: Full reasoning provided ### 4. Email Generator Agent - Role: Create personalized candidate communications - Input: Score, strengths, recommendation - Output: Context-aware email (invite/waitlist/decline) - Personalization: High (mentions specific strengths)