Skip to content
View gnana70's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report gnana70

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
gnana70/README.md

Hi, I'm Gnana Prasath

Lead Data Scientist | Deep Learning, Generative AI, Computer Vision, NLP

LinkedIn Email GitHub


πŸ§‘β€πŸ’» About Me

  • πŸ’Ό Lead Data Scientist at Yubi, Chennai, India
  • 🎯 9+ years building scalable, end-to-end deep learning solutions in computer vision and NLP
  • πŸ€– Passionate about Generative AI, LLMs, Document AI, and deploying real-world ML products
  • πŸ† Proven track record delivering ML and GenAI projects under tight deadlines through strategic planning and cross-functional collaboration

πŸš€ Skills & Tech

  • Languages: Python
  • Backend: FastAPI, Django Rest Framework
  • Databases: PostgreSQL
  • Cloud Platforms: AWS, Azure
  • Expertise: OCR, Document AI, Generative AI, LLMs, Retrieval-Augmented Generation (RAG), Deep Learning, NLP

🏒 Work Highlights

Yubi (Lead Data Scientist, 2022–Present)

  • Led all Computer Vision initiatives, projects & POCs for Yubi Group
  • Developed a template-agnostic bank statement extraction system (deployed across India, Sri Lanka, Middle East)
  • Fine-tuned Llama 3.2 1B for entity extraction in financial documents
  • Built a custom OCR solution saving $10K+/month compared to AWS Textract
  • Designed a KYC extraction system (7 deep learning models, >99% accuracy, <0.5 sec latency, half a million documents/day, $0.5M/year cost saving)
  • Engineered a scalable form recognition engine (Vision-Language Transformer/Donut, >90% accuracy, 20K+ PDFs/day)
  • Delivered a CPU-only captcha resolver (5+ lakh captchas/day)

TheMathCompany (Senior Associate - Data Science)

  • Led a 10-member team for payment recommendation in the auto industry
  • Built price elasticity and optimization frameworks, Airflow ML pipelines

Accenture (App Dev Senior Analyst)

  • Engineered ML-powered resource planners, classifier chains, Flask APIs
  • Recognized as "Automation Prime certified in AI & Analytics"

🌐 Open Source & Community

  • πŸ… Built a state-of-the-art Tamil Text Recognition model (360Β° natural scene reading, outperforms Tesseract, PaddleOCR, EasyOCR)
  • πŸ“„ Published: "Pharmaceutical inspection using machine vision" (Journal of Advanced Research in Dynamical and Control Systems, Jan 2017)

πŸ“š Education

  • B.E. Mechatronics, Kumaraguru College of Technology (2012–2016)
  • Higher Secondary (Science, 95%), SRC Memorial Matriculation School

πŸ’¬ Let's Connect!


β€œTurning ideas into scalable AI solutions.”

GitHub stats

Pinned Loading

  1. tamil_ocr tamil_ocr Public

    OCR Tamil is a powerful tool that can detect and recognize text in Tamil images with high accuracy on Natural Scenes

    Python 68 14