SkillMatch is a premium, intelligent internal tool designed to revolutionize the recruitment process. By leveraging advanced Natural Language Processing (NLP) and a modern high-performance micro-stack, SkillMatch accurately aligns candidate resumes with job descriptions, revealing hidden gems and surface-level misalignments with surgical precision.
SkillMatch follows a decoupled, modern architecture designed for speed and scalability.
graph TD
subgraph "Frontend Layer (React)"
U[User] --> FE[Vite + React SPA]
FE --> FM[Framer Motion Animations]
FE --> TW[Tailwind UI]
end
subgraph "Logic Layer (FastAPI)"
FE -- REST API --> BE[FastAPI Backend]
BE --> AR[Auth Router]
BE --> DR[Document Router]
end
subgraph "Neural Engine (NLP)"
DR --> IE[Inference Engine]
IE --> SE[Skill Extractor]
IE --> MG[Matching Engine]
IE --> SGE[Suggestion Engine]
end
subgraph "Data Layer (MongoDB)"
BE --> DB[(MongoDB)]
DB --> US[User Settings]
DB --> DH[Document History]
end
The "Neural Pulse" of SkillMatch lies in its multi-stage NLP pipeline:
- Ingestion & Preprocessing: Raw text is extracted from PDFs/DOCX and normalized (case folding, stop-word removal).
- Entity Recognition: The
Skill Extractoruses high-precision Regex patterns with word-boundary awareness to identify technical and soft skills. - Semantic Match Proxy: Calculates overlap and term frequency to produce a robust match percentage without heavy model overhead.
- Gap Analysis: Missing skills are identified by comparing extracted sets and categorized via a curated Skill Taxonomy.
- Intelligent Suggestions:
- Co-occurrence: Suggests tools often used together (e.g., Python → Pandas).
- Roadmap Generation: Maps missing skills to actionable learning resources.
- Neural Analysis: High-precision matching using TF-IDF and Cosine Similarity.
- Skill Gap Intelligence: Visualize exactly what's missing between a candidate's profile and the job requirements.
- Smart Suggestions: AI-driven recommendations for upskilling or related technologies.
- Journey History: Full persistence of previous analyses for recruiters to track candidate evolution.
- Admin Console: Real-time usage statistics and dynamic skill taxonomy management.
- Zen UX: A premium, responsive interface featuring glassmorphism and smooth micro-animations.
- Frontend: React 19, Vite, Tailwind CSS 4, Framer Motion, Lucide Icons.
- Backend: FastAPI, Uvicorn, Jose (JWT), Pydantic.
- Database: MongoDB.
- NLP Engine: Lightweight Regex-based extraction, Custom Suggestion Engine.
SkillMatch/
├── backend/ # FastAPI Application
│ ├── routers/ # API Endpoints (Auth, Documents)
│ ├── main.py # Backend Entry Point
│ ├── models.py # Pydantic Schemas
│ └── database.py # MongoDB Connection & Config
├── frontend-react/ # Vite + React Application
│ ├── src/
│ │ ├── components/ # Atomic UI Elements
│ │ ├── pages/ # Dashboard, History, Admin, Landing
│ │ └── context/ # Global State Management
├── nlp/ # Core AI Inference Engine
│ ├── matcher.py # Similarity & Gap Analysis
│ ├── skill_extractor.py # Regex Entity Recognition
│ └── suggestion_engine.py # Co-occurrence & Embedding Logic
├── app_data/ # Knowledge Base
│ ├── skills_list.txt # Base Extraction Dictionary
│ └── skill_taxonomy.py # Categorization Logic
├── requirements.txt # Python Dependencies
└── .env # Environment Configuration- Clone the repository.
- Ensure MongoDB is running (local or Atlas).
- Configure the
.envfile:MONGO_URI=mongodb://localhost:27017/ SECRET_KEY=yoursecretkey ALGORITHM=HS256
# Activate virtual environment
python -m venv venv
.\venv\Scripts\activate
# Install dependencies
pip install -r requirements.txt
# Start Server
uvicorn backend.main:app --reloadAPI Docs available at: http://localhost:8000/docs
cd frontend-react
npm install
npm run devApp available at: http://localhost:5173
© 2026 SkillMatch Team. Licensed under MIT.