Ashish Bindra

Ashish
Bindra

Python developer working on OCR, object detection and document translation.

Projects

InvoiceOCR V2

  • Python
  • FastAPI
  • YOLO
  • Tesseract
  • deep-translator
  • pdf2docx
  • OpenCV
  • Built software that translates Dutch invoices (images and PDFs) into English.
  • Added PDF tools for merging, splitting and PDF-to-image conversion.
  • Extracted selectable text directly from PDFs, and used YOLO with Tesseract OCR for scanned pages.
  • Used deep-translator for accurate Dutch-to-English translation.
  • Trained the model on 10,000 invoices over two weeks.
  • Added structured logging for easier debugging and tracking.

PDF2PDF

  • Python
  • Flask
  • pytesseract
  • pdf2docx
  • docx2pdf
  • Translated Dutch PDFs into English.
  • Kept the original table formatting intact in the translated PDF.

Experience

Junior Data Scientist

Blueaves E-Health Pvt. Ltd.

  1. Fisca-quest Python, Flask, Angular, MongoDB
    • Scraped job listings with BeautifulSoup and Selenium.
    • Added a scheduler to run scraping jobs at set times.
    • Built a dashboard where admins can view, search and filter job listings.
  2. Chatbot FastAPI, MongoDB, Rasa
    • Labeled 4,000+ words of training text with intents and entities.
    • Built the website with FastAPI and Jinja, including a chatbot that answers client FAQs.
    • Built an admin panel to add NLU data, train models and change the pipeline.
    • Deployed with Docker Compose.
  3. Invoice OCR FastAPI, YOLOv4
    • Labeled around 5,000 images with LabelImg.
    • Built a web API and web app with FastAPI that serves results from a retrained YOLO model.
    • Performed image segmentation with YOLOv4.
    • Deployed with Docker.

Skills

Languages & databases

  • Python
  • SQL
  • Java
  • MongoDB

Frameworks & tools

  • FastAPI
  • Flask
  • Django REST Framework
  • Django
  • pandas
  • Git