Ixaya - Humble
May 2019 - present
- Management of two main projects: "Data extraction from documents" and "Data extraction from web scraping"
- Used Deep Learning and Image Processing techniques for document classification, image segmentation and text extraction
- Created an API using Flask, AWS Lambda and EC2 instances in order to upscale production depending on requests
- Used Selenium to create bots in order to download or extract data depending on the database missing information and CRUD database handling
- Reduced form filling time from 10 min to 25 sec (in average) by allowing users to upload documents and allows database to maintain itself by extracting data from specific insurance companies