Manuj Rai

A direct-to-consumer storefront with product catalog, cart, checkout, inventory, and an admin panel the owner actually uses.

• Replaced 3 separate tools (sheet + DMs + UPI) with a single product
• Checkout completion lifted noticeably vs the abandoned DM-and-Google-Forms flow it replaced

A customizable salon-booking template — each owner gets a branded deployment with their own services, stylists, slots, payments, and SMS confirmations.

• No-show rate dropped once deposit-on-booking + SMS reminders went live
• Daily scheduling time for the owner went from hours of WhatsApp triage to minutes in the dashboard

An automated newsroom agent that reproduces an editorial team's morning routine — searches the web and ~52 curated sources across 11 countries for textile/apparel news, scores and de-dupes the best stories with an LLM, confirms each against its primary source, then drafts publish-ready articles. Two-phase with a human in the loop, streamed live over SSE.

• Turns the desk's manual morning scan into one click — a deduped shortlist of ~20–30 ready-to-draft story candidates
• Two-phase pipeline keeps cost low: a scan is ~4 LLM calls / a few thousand tokens on gpt-4o-mini, and only editor-picked stories get the expensive full fetch

A production retrieval-augmented chatbot over internal documents. Metadata-filtered retrieval feeds a tool-using agent loop; FastAPI exposes the API; the Python service runs on Render.

• Metadata filtering narrows the candidate set before vector ranking — fewer off-topic chunks reach the LLM
• Agent with tools handles structured questions ("extract X from section Y") beyond plain semantic search