Subhajit Mandal

I’m an AI Engineer with 4+ years of experience architecting production-grade AI systems—LLM-powered applications, autonomous AI agents, and end-to-end AI business automation. I focus on eliminating manual workflows through intelligent agent-to-API orchestration and measurable operational impact.

I specialize in building RAG pipelines, Voice AI integrations (Vapi, OpenAI Whisper), and scalable async backends using FastAPI and LLM APIs (OpenAI/Anthropic). I’ve deployed GenAI microservices and built prompt engineering and model evaluation frameworks that support reliable production behavior.

In my recent role, I architected and deployed a production Voice AI Agent integrating Vapi with Dentally CRM via FastAPI, Webhooks, and MCP—enabling fully automated 24/7 appointment booking with zero manual intervention. I implemented secure HMAC webhook verification, real-time REST API workflows, and conversation-to-action mapping that reduced booking errors and removed 100% of manual scheduling overhead.

Previously at Mphasis, I built and deployed Python-based backend services and REST APIs (Flask) for intelligent automation and AI-ready workflows in the BFSI domain, including SQL optimizations that improved query execution by 40%. At Amazon, I used data analysis on customer complaints to deliver actionable stakeholder insights that reduced complaint resolution time.