AI Powered - Dark Store Operations Loss Attribution & Efficiency Analysis
Developed an AI-powered Streamlit application using Ollama (Llama 3.2) to generate executive insights and prioritized recommendations from verified analytical findings Dark-store operations involve high-velocity inventory movement across receiving, storage, picking, packing, and dispatch. Operational incidents such as handling errors, picking mistakes, expiry, packaging issues, product defects, and storage-related problems can create recurring financial losses.
Overview
Developed an AI-powered Streamlit application using Ollama (Llama 3.2) to generate executive insights and prioritized recommendations from verified analytical findings Dark-store operations involve high-velocity inventory movement across receiving, storage, picking, packing, and dispatch. Operational incidents such as handling errors, picking mistakes, expiry, packaging issues, product defects, and storage-related problems can create recurring financial losses.
Business value
- Built KPI dashboard.
- Analyzed Loss Drivers
- Generated business recommendations.
Technologies
- Power BI
- python
- SQL
- KPIs Engineering
- Feature Engineering
- Dashborad
- Excel
- DAX
- ollama (for AI genreated Insights)
Problem, solution & architecture
Detailed problem framing, solution design, architecture diagrams and challenge notes for this project aren’t published yet. The repository linked above documents the data model, transformation steps and analysis notebooks in full.