AWS KYC Verification System

AI-assisted KYC verification system using AWS Rekognition, Textract, Lambda, and S3.

Problem Statement

Manual KYC verification is slow, error-prone, and vulnerable to fraud. This project demonstrates a serverless, AI-assisted KYC pipeline that verifies identity consistency using face comparison and document text extraction.

What This System Does

• Captures selfie and ID document images
• Verifies face similarity between selfie and document
• Extracts document text using OCR
• Applies rule-based decision logic (PASS / FAIL)

What This System Does NOT Do

• Does NOT perform forensic document authenticity checks
• Does NOT detect holograms or security threads
• Does NOT store user images permanently

Architecture Overview

• Frontend uploads images using pre-signed S3 URLs
• AWS Lambda processes verification
• AWS Rekognition handles face detection and comparison
• AWS Textract extracts document text
• Decision engine determines verification status

AWS Services Used

• Amazon S3
• AWS Lambda
• Amazon API Gateway
• AWS Rekognition
• AWS Textract
• Amazon DynamoDB (results only)

Security Considerations

• Private S3 buckets with encryption
• Short-lived pre-signed URLs
• IAM least privilege
• Automatic image deletion

Project Status

Phase 1 – MVP in progress