Projects
AI-Powered Payment & Sales OCR
Python · Google Gemini AI · gspread · watchdog
​​
The Problem
​
Handwritten payment receipts, sales invoices, and credit notes were being manually read and entered into ledgers across multiple Google Sheets — a slow, error-prone process that happened dozens of times a day.
​
The Solution
​
Built a real-time folder watcher that detects when a receipt photo is dropped into a folder, sends it to Google Gemini AI for extraction, identifies the customer and brand automatically, checks for duplicates, and posts the data directly to the correct ledger tab in Google Sheets.
The Result
​
Zero manual data entry for payments and sales. Receipts are processed and posted to live ledgers within seconds of being photographed. Duplicate prevention and multi-brand routing built in.
Sales Automator
Python · pandas · openpyxl
​
The Problem
​
Every month, our sales team spent 7 full working days manually consolidating sales data from hundreds of dealer files into a single master report. The process was error-prone, exhausting, and left no time for actual analysis.
​
The Solution
​
Built a Python automation tool that pulls data from all dealer files, applies business rules, validates entries, and generates a clean, formatted master sales report — automatically.
The Result
​
7 days of manual work now completes in under 2 minutes. Zero copy-paste errors. The sales team gets their report before their morning coffee.
Purchase Processor
The Problem
Processing purchase data across a large and growing number of supplier files was a daily bottleneck. Each file had to be opened, validated, and entered manually — a process that scaled badly as the business grew.
​
The Solution
​
Built a tool that automatically reads, validates, and processes purchase data across any number of input files — regardless of volume. No manual opening, no manual entry.
​
The Result
​
Purchase processing time cut to a fraction of the original. The tool handles 1 file or 1,000 files with equal speed, making it fully scalable as the business grows.
Daily Data Splitter
The Problem
Every day, a consolidated data file had to be manually split and distributed across 314 individual dealer files. This meant hours of repetitive sorting, copying, and pasting — every single day.
The Solution
​
Built an automated splitter that reads the master daily file, identifies each dealer's data, and writes it to the correct dealer file — all in one run.
The Result
What used to take hours now runs in minutes, every day. 314 dealer files updated automatically with zero manual sorting.
Sales Report Generator
The Problem
Generating formatted, shareable sales reports required manually pulling data, arranging it across sheets, applying formatting, and double-checking figures — a slow process prone to inconsistency.
The Solution
Built a report generator that pulls live sales data, applies formatting rules, and outputs a professionally structured Excel report on demand — ready to share immediately.
The Result
Reports that used to take hours to prepare are now generated in seconds, with consistent formatting every time. No formulas to check, no formatting to fix.
AI Powered Physical Inventory Auditor
The Problem
Auditing physical eyewear inventory meant manually reading hundreds of product labels — recording model numbers, color codes, sizes, MRP, and manufacturing dates by hand into a spreadsheet. Time-consuming, tedious, and error-prone.
The Solution
Built an AI-powered tool that processes photos of product labels using Google Gemini AI. It automatically extracts brand, model number, color code, size, MRP, and manufacturing date from each label — detecting brands like Ray-Ban, Vogue, Oakley, Versace, and Burberry automatically from model number prefixes. Outputs a formatted Excel inventory sheet with a full summary.
The Result
A stack of product photos becomes a complete, formatted inventory report in minutes. Manual label reading eliminated entirely.
