🚀 HireLens – AI-Powered Job Application Optimizer + Semantic Ranking (Fullstack-project,.NET + React)

HireLens is an enterprise-grade AI recruitment platform designed to modernize hiring through intelligent automation, AI-powered evaluation, and transparent decision-making.

---

🌟 Vision

HireLens aims to redefine modern recruitment by combining intelligent automation, AI-driven evaluation, and transparent decision-making into one seamless platform.

The mission is to empower recruiters to make faster, smarter, data-driven hiring decisions while ensuring candidates receive fair, evidence-based evaluations.

By integrating AI scoring, explainability, and structured workflows, HireLens bridges the gap between human intuition and machine intelligence in the hiring ecosystem.

---

🌍 Live Demo

Frontend (Vercel):
https://ai-powered-job-application-optimizer-vupphwx74.vercel.app

Backend API (Render):
https://ai-powered-job-application-optimizer.onrender.com/swagger

---

🏗 Architecture

Frontend (Vercel)
⬇
Backend API (Render – Dockerized)
⬇
PostgreSQL Database (Render Managed DB)

🧩 System Design Focus

This project demonstrates full-stack architecture, API design, and scalable backend development using modern technologies.

Frontend (React)

   │
   ▼

Backend API (.NET Core)

   │
   ▼

Database (PostgreSQL)

   │
   ▼

AI Service (OpenAI)

---

🛠 Tech Stack

Backend

• ASP.NET Core 8
• Entity Framework Core
• PostgreSQL (Npgsql)
• JWT Authentication
• Swagger

Frontend

• React (Vite)
• Axios
• JWT Token Storage

AI

• OpenAI (Evaluation + Embeddings)

Infrastructure

• Render (API + Database)
• Vercel (Frontend)
• GitHub Actions (CI/CD)
• Docker

---

✨ Features

🔐 Authentication

• JWT-based authentication
• Secure password hashing
• Role-based access ready

👤 Candidate Management

• Create candidate
• Update status (New → Shortlisted → Interview → Hired → Rejected)
• Recruiter notes
• Audit logging

📄 Resume Upload

• PDF-only upload (15MB max)
• Text extraction using PdfPig
• Cleaned text storage
• Ready for AI evaluation

🤖 AI Evaluation

• OpenAI integration
• Skill extraction
• Match scoring
• Embedding generation
• Explainability service

📊 Ranking Engine

• Score-based candidate ranking
• Evaluation tracking

🧾 Audit Trail

Tracks:

• Who changed status
• Before/After snapshot
• Timestamp

---

🐳 Docker

Build and run backend container:

docker build -t hirelens-api .
docker run -p 8080:8080 hirelens-api

---

⚙️ Environment Variables (Production)

Render backend requires:

ConnectionStrings__DefaultConnection=Host=...;Port=5432;Database=...;Username=...;Password=...;SSL Mode=Require;Trust Server Certificate=true;

Jwt__Key=VERY_LONG_SECRET_32+_CHARS
Jwt__Issuer=HireLens
Jwt__Audience=HireLens

OpenAI__ApiKey=sk-...
OpenAI__Model=gpt-4o-mini
OpenAI__EmbeddingModel=text-embedding-3-small

---

🧪 CI/CD

GitHub Actions pipeline includes:

• .NET Restore
• Build
• Tests
• Frontend Build

---

💻 Local Setup (Without Docker)

Backend

Update HireLens.Api/appsettings.json:

• ConnectionStrings:DefaultConnection
• Jwt:Key (32+ characters)
• OpenAI:ApiKey (optional)

Run migrations:

dotnet ef database update

Run API:

dotnet run --project HireLens.Api

Swagger:

http://localhost:5159/swagger

---

Frontend

cd HireLens.Web
npm install
npm run dev

---

🐳 Docker Compose

From repository root:

docker compose up --build

Web:

http://localhost:3000

API:

http://localhost:8080/swagger

---

🔑 Key Endpoints

POST /api/auth/register
POST /api/auth/login
GET  /api/candidates
POST /api/resumes/upload
POST /api/evaluations/jobs
POST /api/evaluations
GET  /api/rankings/jobs/{jobId}/candidates
GET  /api/rankings/jobs/{jobId}/candidates/{candidateId}/explain

---

🧪 Tests

dotnet test

---

📝 Notes

• If OpenAI key is not configured, evaluation endpoints return a short 400 error.
• PDF extraction works for text-based PDFs. Scanned/image PDFs require OCR.

---

📌 Future Improvements

• Role-based authorization (Admin / Recruiter)
• Resume similarity search
• Vector database integration
• Analytics dashboard
• AI interview question generator
• Production-grade logging (Serilog)
• Rate limiting

---

👨‍💻 Author

Thabang Rakeng
Full-Stack Developer | AI-Focused Backend Engineer