/Team-45

Primary LanguageJupyter Notebook

Reducing the lead time and error rate in KYC verification process at financial institutions

  • Our project aims to automate the KYC process for financial institutions, namely Banks, Non-Banking Financial Companies (NBFCs), and Payment Aggregators, by extracting data from customer submitted documents at application stage.
  • We aim to provide an INDUSTRY READY solution where the user uploads the required documents
  • We are eliminating the human dependence and manual intervention for submitted document’s verification.
  • Any anomaly noted through document verification will be informed to the user at application stage itself.
  • Turn around time of document verification will be reduced from days to not more than 30 seconds.

    How are we solving this problem?

  • Using machine learning algorithms, we aim to automate processing of submitted customer personal information documents for KYC norms.
  • Using OCR, NLP and CV technologies, meaningful data can be extracted from uploaded documents.
  • Extracted data is processed further and matched to customer declared details through which an anomaly report is generated.
  • Known anomalies in submitted documents are communicated on real-time basis to uploader( consumer) and financial institutions.
  • Onus of completing KYC timely with correct documents will shift entirely to customers.

    Tech Stack Used:

  • React JS : React will be used to design the website
  • Tailwind CSS : Tailwind CSS is used to provide styling to the website
  • FastAPI : FastAPI provides an intuitive and easy-to-use interface for building scalable and async web applications with automatic interactive API documentation and validation.
  • Computer Vision : This field of AI will be used for image classification and image processing
  • OCR : Optical Character Recognition will be used for extracting text from the documents
  • Deep Learning : Deep Learning will be used to create state of the art classification models for aur custom image classification problem.