How d102-preview/preview works
High-Level Components
The system consists of three main services that should be prominently displayed:
a) Frontend Service
• React-based web application
• As the client-facing layer
• Features: Resume Upload, Interview Practice, Real Interview, Analysis Results
b) Backend Services (Split into two parts)
• Main Backend (Spring Boot)
• User Management
• Interview Management
• Resume Processing
• Question Management
• AI Backend (FastAPI)
• Face Analysis
• Answer Analysis
• Question Generation
c) Data Stores
• MariaDB (Main database)
• Redis (Cache and session management)
• File Storage (For resumes and recordings)
Component Relationships and Data Flow
• Frontend → Backend: REST API calls
• Backend → AI Service: Internal API calls
• Backend ↔ Databases: Data persistence
• AI Service ↔ ML Models: Model inference
External Services Integration
Connections to:
• OpenAI API (for question generation)
• FastAI API (for face analysis)
• Email Service
Infrastructure Layer
Deployment infrastructure:
• Docker containers
• Jenkins pipeline
• Nginx reverse proxy
• SSL/TLS termination
Specific Implementation Details
Technical stack indicators:
• Frontend: React, TypeScript, Tailwind CSS
• Backend: Spring Boot, JPA, QueryDSL
• AI: FastAPI, PyTorch, KoBERT
• Infrastructure: Docker, Jenkins, Nginx
Security Layer
Security components:
• JWT authentication
• API Gateway
• Security filters
Infrastructure Layer
Key Design Considerations:
• Microservices architecture
• Clear separation between AI and main backend services
• Real-time features (SSE for interview feedback)
• Monitoring and logging components
• Scalability points