Hindustan Bills

Inspiration

Anyone who has stood in a long billing queue at a grocery store understands the frustration—especially when digital payments exist but checkout still feels slow and outdated. The inspiration for Hindustan Bills came from observing how small and mid-scale retailers in India struggle with inefficient billing systems while also having zero access to data-driven insights that large retailers take for granted. I wanted to build something that: Improves customer experience instantly Requires no expensive hardware Helps retailers make smarter decisions using their own data This project was born at the intersection of retail pain points, backend engineering, and applied machine learning.

What it does

Hindustan Bills is a queue-less smart billing system. Customers scan product barcodes using their phone Items are added to a virtual cart in real time Payments are completed digitally Customers proceed to a final verification counter and exit Behind the scenes, every transaction becomes structured data that powers ML-driven analytics for retailers—transforming billing into intelligence.

How we built it

The system is designed with a production-first mindset, not just a demo. Backend Architecture: RESTful APIs built using Node.js and Express JWT-based authentication for secure user access Modular controllers for users, products, carts, and orders MongoDB for flexible, scalable data storage

Checkout Flow: User authentication Product scanning & cart creation Order placement Payment simulation Store-side verification

Machine Learning & Analytics (Core Innovation): Instead of treating ML as an afterthought, the data model was designed from day one to support analytics. Planned and partially implemented ML capabilities include: Demand forecasting using historical sales data Low-stock prediction for inventory management Sales trend analysis (time-based & product-based) Market basket analysis to detect frequently bought item pairs These insights enable retailers to reduce stock-outs, avoid overstocking, and optimize store operations.

Challenges we ran into

Designing a secure yet fast checkout flow without physical POS hardware Structuring transaction data to be ML-ready Balancing simplicity for customers with control for retailers Ensuring the system is scalable beyond a single store

Each challenge helped refine both the technical design and the business logic of the platform.