ShelfScan

Retail, VisionAgent

I built ShefScan, a Vision AI powered tool that redefines how retailers track and manage shelf inventory. With just a photo of store shelves, ShefScan delivers real time product counts and visibility insights, eliminating the need for manual stock checks.

I built the solution using LangChain, Vision AI, and LLMs, combining advanced computer vision with intelligent automation to achieve high accuracy, even in complex retail environments. By reducing human effort, ShefScan improves stock accuracy, minimizes errors, and boosts operational efficiency. By blending AI-driven precision with user friendly design, I enabled supermarkets and retail chains to make smarter, data backed decisions for shelf management.

With ShefScan, retailers can quickly spot missing items, optimize shelf space, and improve product placement, directly impacting sales opportunities. Within weeks of deployment, businesses reported faster audits, better visibility, and reduced stock monitoring costs.

Tools Used

  • Visual Studio Code
  • Slack
  • Streamlit
  • Docker
  • Vision Agent
  • LangChain
  • OpenAI
  • GPT
  • Gemini AI
  • Claude
  • HuggingFace
  • LLM
  • Anthropic
  • Vertex AI