SuperCoder AI Interview Platform

Overview

As Senior Software Engineer at SuperCoder, I architected and developed an enterprise-grade AI interview platform that automates technical hiring through intelligent candidate assessment, handling 10,000+ monthly interviews across 100+ organizations.

My Role & Contributions

• Architected microservices ecosystem with 5 specialized backend services
• Built AI interview engine using Azure OpenAI GPT-4 for intelligent assessments
• Implemented WebRTC video platform supporting 500+ concurrent sessions
• Designed multi-tenant PostgreSQL architecture with dynamic schema switching
• Led integration of Stripe payments processing $500K+ subscriptions

Technical Architecture

Microservices Design

• Recruiter Service: Java/Spring Boot core with JWT authentication
• AI Service: Python/FastAPI leveraging Azure OpenAI
• Candidate Service: Automated onboarding and resume parsing
• Payment Service: Stripe integration with webhook processing

AI Innovation

• Dynamic question generation based on candidate background
• Real-time conversational interviews with follow-ups
• Token optimization for reduced API costs
• High interview completion rate with fallback mechanisms

Real-time Communication

• Agora WebRTC with HD video/audio streaming
• Cloud recording with automatic chunking
• Real-time proctoring and violation detection
• Audio waveform visualization using Web Audio API

Key Achievements

Technical Challenges Solved

• Multi-tenant Isolation: Dynamic schema switching with complete data separation
• WebRTC Connectivity: Cross-firewall video streaming using TURN servers
• AI Latency: Optimized prompts and response streaming
• Scale: Stateless services enabling horizontal scaling

Tech Stack