Hi, I'm Hemant

> Led a 3-member team delivering 3 core products on Healos: AI Scribe, AI Voice Agent, and an Email Campaigning Platform using OpenAI GPT, Gemini API, and LangChain

> Built real-time speech-to-text pipelines with Whisper and Deepgram, plus RAG pipelines with pgvector for context-aware note retrieval and semantic search across medical knowledge bases

> Engineered multi-step prompt chains for accurate, structured clinical summaries and voice-driven workflows

> Designed a secure multi-tenant system with complete data isolation across clinics and clients, improving user engagement by 25%, reducing API response time by 25%, and decreasing UI bugs by 20%

> Improved high-load system performance by 40%, reduced script execution time by 45% through advanced PostgreSQL query optimization, and containerized the full stack with Docker

> Deployed and managed API servers on GCP using VMs and Container Apps with Nginx load routing, while implementing Grafana monitoring, logging, and database backup strategies

> Boosted system performance by 40% under load by implementing optimized RESTful APIs, refactored database models, and improved request handling on a Node.js/Express backend

> Cut API response times by 25% through scalable backend architecture improvements and strategic query optimization

> Improved user management efficiency by 30% by streamlining core portal workflows and admin capabilities on the internal ops platform

> Delivered responsive React.js UI components that lifted user engagement by 35% and reduced UI-related bugs by 20%

> Automated 50% of manual internal workflows by collaborating with cross-functional teams to upgrade tooling and expand admin functionality, significantly reducing operational overhead

> Delivered a responsive frontend, applied code refactoring, and implemented best practices and design patterns to ensure a robust and maintainable structure

> Drove a 15% increase in user adoption through improved usability and cleaner product flows

> Cut script execution time by 50% and improved user management efficiency by 30%, enhancing overall platform performance