AI Lead Scoring Service
Built an AI-powered lead scoring backend achieving 90% prediction accuracy, reducing manual qualification time by 70%. Developed an asynchronous Spring WebFlux backend processing 10,000+ leads per batch using OpenAI API integration. Implemented a CSV ingestion pipeline handling 6+ lead attributes and a hybrid scoring engine combining rule-based logic with GPT-powered reasoning, achieving sub-1.5s response time. Deployed to Railway with CI/CD via GitHub Actions, implementing secure environment variable management.
// INTERVIEW ENGINE