Atin Sharma

I’m a Full-Stack Product Engineer with founding-team experience shipping AI-powered products for Indian SMBs. At OpenBiz, I architected the core MERN platform from scratch and scaled it to 1,000+ active users using TypeScript for type safety and maintainability.

I built Vyapar—a WhatsApp-native LLM assistant—covering the full message lifecycle from webhook receipt to LLM processing and response dispatch. I diagnosed and resolved N+1 query patterns in Supabase (PostgreSQL), implemented indexing and response caching, and cut API latency by 40% across critical endpoints.

Beyond assistants, I engineered real-time bidirectional chat with Socket.IO and integrated VideoSDK for in-app conferencing. I also integrated Razorpay subscription billing end-to-end (webhook validation, payment state reconciliation, failed-payment retries) and owned CI/CD with GitHub Actions and Vercel, reducing deployment cycle time by 30%.

I’m especially focused on product engineering for conversational AI, document intelligence, and compliance automation. I’ve built RAG systems like Nimbus (pgvector + Groq streaming, provider-agnostic LLM switching) and DocGPT-RAG (PDF ingestion with PyPDF2 chunking, ChromaDB embeddings via all-MiniLM-L6-v2, and a Django REST orchestration backend) to make retrieval accurate, cost-aware, and production-ready.