JobVyn

A distributed recruitment platform built with a service-oriented architecture, subscription-based access control, and AI-powered utilities.

Designed to demonstrate real-world backend engineering: API gateway, authentication, relational data modeling, event-driven workflows, payment verification, and containerized deployment with GitHub Actions CI/CD.

---

---

Overview

JobVyn is a multi-service recruitment system where Recruiters post jobs and Candidates apply through a subscription model.

The architecture separates concerns into independent services communicating via REST and Kafka, with a central API Gateway handling all inbound traffic. This simulates production-grade backend systems rather than a monolithic demo project.

Core focus areas:

• Central API gateway with rate limiting and request tracing
• Clean service boundaries with isolated deployments
• Secure role-based access control
• Event-driven background processing
• Structured AI integration
• Cloud-ready containerized deployment

---

Tech Stack

Backend

Layer	Technology
Runtime	Node.js + Express
Database	PostgreSQL
Message Broker	Kafka
Gateway Cache	Redis
Auth	JWT
Payments	Razorpay
AI	Gemini API
Containerization	Docker
Hosting	AWS EC2
CI/CD	GitHub Actions

Frontend

Layer	Technology
Framework	Next.js + TypeScript
Styling	Tailwind CSS + shadcn/ui

---

Architecture

All client requests enter through a single API Gateway (port 8001) which proxies them to the appropriate downstream service.

Client
  │
  ▼
API Gateway (:8001)
  ├── /api/auth     →  Auth Service    (:5001)
  ├── /api/user     →  User Service    (:5002)
  ├── /api/job      →  Job Service     (:5003)
  ├── /api/utils    →  Utils Service   (:5004)
  └── /api/payment  →  Payment Service (:5005)

Services

Service	Responsibility
API Gateway	Central entry point — routing, rate limiting, request tracing
Auth Service	Authentication, JWT issuance, role management
User Service	Profiles, subscriptions, job applications
Job Service	Company creation, job lifecycle management
Payment Service	Payment verification, subscription activation
Utility Service	Async event processing, AI-powered features
Kafka Broker	Non-blocking background workflows (e.g., email notifications)

Each service is isolated, containerized, and environment-configured for deployment.

---

API Gateway

The gateway is the single entry point for the entire platform. It handles cross-cutting concerns so individual services stay focused on business logic.

Features

Request Tracing

• Assigns a unique x-request-id (UUID v4) to every request
• Propagates the ID downstream to all services
• Logs request start, response status, and total duration per request

Redis-based Rate Limiting

• IP-based sliding window strategy
• Default: 100 requests per 60-second window
• Returns HTTP 429 with X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset headers
• Graceful fallback — if Redis is unavailable, requests pass through

Reverse Proxy

• Routes traffic to 5 downstream services via http-proxy-middleware
• 10-second proxy timeout
• Stamps outgoing requests with x-gateway-service: gateway
• Stamps responses with x-proxied-service: {serviceName}
• Forwards X-Forwarded-* headers for accurate IP detection

Centralized Error Handling

• 502 Bad Gateway on proxy failures with socket cleanup
• 404 Not Found for unmatched routes
• 500 Internal Server Error for unhandled exceptions

Middleware Stack (in order)

requestIdMiddleware  →  loggerMiddleware  →  rateLimitMiddleware
     →  serviceProxy  →  notFoundMiddleware  →  errorMiddleware

---

Key Capabilities

• Role-based authentication (Recruiter / Candidate)
• Subscription-gated job applications
• Company and job management workflows
• Duplicate application prevention
• Event-driven email notifications via Kafka
• AI-powered resume analysis (structured output)
• AI-based career guidance generation
• Dockerized multi-service deployment on AWS EC2
• Automated CI/CD pipeline via GitHub Actions

---

System Flow

1. Request hits the API Gateway
2. Gateway assigns a request ID and logs the request
3. Rate limiter checks the IP against Redis
4. Request is proxied to the target service
5. Response is logged with status and duration

Application flow:

1. User authenticates → receives JWT
2. Recruiter posts jobs under a company
3. Candidate activates subscription via Razorpay
4. Application triggers a Kafka event
5. Utility service processes the event asynchronously
6. AI endpoints provide resume scoring and recommendations

---

Design Principles

• Single entry point via API Gateway for all traffic
• Clear separation of concerns across services
• Server-side validation of all business rules
• Asynchronous processing for scalability
• Structured AI responses for predictable frontend rendering
• Production-oriented deployment strategy

---

Project Structure

Each service follows a layered structure:

routes → controllers → business logic → database access

The gateway follows:

middleware → proxy factory → service config → error handling

---

Status

• Backend services complete
• API Gateway with Redis rate limiting operational
• Kafka integration operational
• AI features integrated
• Dockerized and deployed on AWS EC2
• CI/CD pipeline active via GitHub Actions

---

Arun Kumar Backend Engineer | Distributed Systems Focus