CallSense AI

Details:

• Independently designed and developed an automated call processing system that transcribes phone recordings, extracts booking details, and sends notifications, reducing manual administrative work by 80%.
• Architected a modular NestJS/TypeScript backend with specialized components (PhoneModule, TranscriptionModule, LLMModule) utilizing dependency injection for improved maintainability and testability.
• Engineered an AI-powered processing pipeline utilizing OpenAI technologies that transcribes speech-to-text (Whisper), cleans conversation transcripts, generates concise summaries, and extracts structured booking data with language model analysis (ChatGPT).
• Developed service abstraction layers for multiple external APIs (Telnyx, OpenAI, Postmark) that isolated core business logic from third-party dependencies, simplifying maintenance and future API updates.
• Coded an asynchronous processing system with comprehensive error handling and state persistence, ensuring system reliability and preventing redundant processing of recordings even during service interruptions.