Browse categories
English
$
USD
Connect to your AI
I will extract invoice, receipt and pdf data to excel
Kaushalya
About this gig
ChatGPT can read one document. Your business handles hundreds every week.
Manual data entry from invoices, bills, and receipts drains hours and creates costly accounting errors. I build production-grade document extraction pipelines that turn messy PDFs, scans and photos into clean, validated Excel spreadsheets automatically.
What I extract:
- Invoice #, PO & Order IDs
- Issue & Due Dates (normalized to YYYY-MM-DD)
- Vendor & Client Info (Name, Address, Tax ID)
- Line Items Table (Description, Qty, Unit Price, Line Total)
- Subtotal, Tax, Shipping & Grand Total
- Math Validation (flags discrepancies automatically)
Supported formats:
- Native digital PDFs
- Scanned & multi-page documents
- Phone photos & receipts (PNG, JPG)
Deliverables:
- Clean Excel (.xlsx) / CSV spreadsheets
- Structured JSON data (Pydantic validated)
- Python script, FastAPI endpoint or Streamlit web dashboard
Why work with me?
- Handles multi-column layouts, slanted scans and low-res receipts
- Fast communication and 100% data privacy
Special Offer for Your Peace of Mi
Get to know Kaushalya
Kaushalya
Machine Learning and Python Engineer OCR, LLMs and Document AI
- FromSri Lanka
- Member sinceJul 2026
Languages
Sinhala, English
I'm an Applied Machine Learning Engineer. I help startups, businesses and developers build reliable, production-ready AI solutions that automate workflows and turn raw data into clean business outputs.
My expertise includes:
- Production RAG systems with verified citations & zero hallucination
- Autonomous AI agents with security guardrails & tool calling
- Document AI & OCR pipelines (PDFs, invoices, receipts to Excel/DB)
- Scalable FastAPI backends & NLP classification models
My priorities: reliability, security & clean Python code. Let's discuss your project in the inbox!
My Portfolio
Other AI Development Services I Offer
FAQ
Can you extract data from blurry scans or phone photos of receipts?
Yes. I use pre-processing techniques (contrast enhancement, adaptive thresholding and noise reduction) before OCR to extract text from slanted, faded or phone-captured documents with high accuracy.
What formats will the extracted data be delivered in?
I deliver formatted Excel (.xlsx), CSV or structured JSON files. If you need the extraction code, I can also provide a standalone Python script, a FastAPI endpoint or an interactive web dashboard.
Can I test accuracy on my documents before placing an order?
Absolutely! Message me in the inbox with 1 sample document. I will run a free test extraction and send you the Excel preview so you can verify accuracy before ordering.
Is my document data kept private and confidential?
100% confidential. Your files are processed locally on secure environments, never shared and permanently deleted after delivery. An NDA can be signed if required.
