At DataFlask, I build Python backend infrastructure, agentic AI platforms, RAG workflows, and conversational systems. My work spans product delivery from requirements and solution design through development, testing, deployment, and maintenance.
I engineered an autonomous procure-to-pay platform covering requisitions, budget validation, approvals, purchase orders, supplier dispatch, and ASN tracking. I also built AI agent tools and natural-language workflows with WhatsApp and email approvals, real-time updates, and tamper-evident audit trails.
I built a LangGraph-based customer operations platform with tool calling, conditional workflows, conversational memory, and multi-step task execution. Its RAG implementation uses embeddings, pgvector, and semantic search with source attribution, alongside human-in-the-loop authorization and prompt-injection defenses.
I developed conversational solutions across WhatsApp, Instagram, and Facebook Messenger, including an AI-powered support platform that routed requests across 40+ business intents. Other work includes a WhatsApp-based education platform handling approximately 1,000–10,000 messages per day and O2C and dealer management automations.
