How vectornguyen76/resume-ranking works
Major Components
• Frontend: Built using Next.js with TypeScript and TailwindCSS. It serves as the user interface and dashboard for candidate evaluation, job listings, and AI-powered analytics.
• Backend: Developed with FastAPI (and elements of Flask as noted in the README) and MongoDB. It exposes REST API endpoints for candidate, job, and matching operations. It handles request routing, business logic, data validation, and integration with the analysis service.
• Analysis Service: A separate service (with its own Dockerfile and configurations) whose role is to perform resume and job description analysis using OpenAI GPT models and LangChain. It includes clear modules for candidate, job, and matching, each with its own routing, schema definitions, and prompt configurations.
• Infrastructure & Orchestration: This includes CI/CD pipelines (all GitHub Actions workflows for development, staging, production, and rollback), Docker Compose orchestration (in docker-compose.yml), and Nginx configuration (to act as a reverse proxy) to manage deployment.
• External Services: External dependencies such as OpenAI’s GPT models, AWS (as indicated in the README), and potentially MongoDB hosted on the cloud. These elements are key for data processing, storage, and AI functionalities.
Relationships and Data Flows
• The Frontend communicates with the Backend API based on environment configuration (see the NEXT_PUBLIC_API_URL in the frontend’s .env.production).
• The Backend processes incoming API requests by interacting with its internal controllers, services, and schemas. For operations requiring heavy AI processing (like resume parsing or job description analysis), it forwards the request to the Analysis Service.
• The Analysis Service uses LLMs, custom prompt engineering, and function calling with OpenAI APIs to process resumes or job descriptions, then returns the analyzed data to the Backend.
• The Backend may also interact with a persistence layer (MongoDB) for storing candidate information, analysis results, and job data.
• CI/CD pipelines drive automated testing, builds, and deployments to various environments. Docker containers (Dockerfiles in each service and docker-compose.yml) are used to run and manage the services.
• Nginx is employed to route traffic, manage SSL and reverse-proxy operations, and provide load balancing between services if necessary.
Architectural Patterns and Technologies
• Separation of Concerns: The application divides responsibilities clearly across the frontend, backend, and analysis service.
• Microservices/Modular Architecture: Although managed within one repository, the clear division into separate services means each can be developed, deployed, and scaled independently.
• Containerization: Use of Docker and docker-compose to orchestrate multi-service deployment.
• CI/CD Automation: GitHub Actions workflows to ensure code quality, automated testing, and seamless deployment (with environment-specific pipelines).
• AI Integration: Usage of OpenAI GPT, LangChain, and structured prompt engineering clearly outlined in the README and file structure.
Final Touches
• Relevant key configuration files (e.g., docker-compose.yml, Nginx config) in a “Deployment/Infrastructure” section.
• Be as detailed as possible to capture the granularity—the multiple internal modules (candidate, job, matching) in both the analysis service and backend deserve representation for complete understanding.