Abhinav Prakash

• Custom NER for Name Screening Discounting

- Develop custom NER to identify and extract various attributes (like occupation, employer, nationality, etc.) to enhance HOLMES Name Screening Discounting product

- Uses CRF and LSTM algorithms

• Entity extraction from images

- Develop a pipeline to pre and post process images for text extraction using OCR services(Google Tesseract), for bills and invoices
- Image classification used to classify document types
- Used OpenCV and transfer learning Anomaly detection in audits
- Develop AI pipeline for anomaly detection in highly-imbalanced datasets rom audit processes in an insurance claims system
- Used XGB and CNN for developing models

• Created a Web Scrapping API that automates the search of all background details of a person with the help of Google
• Collaborated with cross-functional teams to integrate data science solutions into existing systems, streamlining operations.
• Assisted in developing a predictive analytics framework that enabled proactive decision-making across departments.