Cloud & AI Case Study
AI Document Intelligence Platform
A serverless document processing platform developed as an AWS AI course project. The solution explored text extraction, entity recognition, sentiment analysis, and AI-assisted summarization for PDF and image documents.
Project Overview
The platform allowed users to upload PDF, PNG, and JPG documents and start an asynchronous processing workflow. Processed results included extracted text, detected entities, sentiment analysis, and AI-assisted summaries that could be reviewed and exported as JSON.
Architecture
The React frontend was delivered through Amazon S3 and CloudFront. Amazon Cognito provided authentication, while API Gateway and Python-based AWS Lambda functions handled application requests. AWS Step Functions coordinated processing across Amazon Textract, Amazon Comprehend, and Amazon Bedrock. Amazon S3 stored uploaded documents, and DynamoDB stored document metadata and analysis results.
React frontend
→ CloudFront and Amazon S3
→ API Gateway and Amazon Cognito
→ AWS Lambda and Step Functions
→ Amazon Textract, Amazon Comprehend, and Amazon Bedrock
→ Amazon S3 and DynamoDB
Key Capabilities
- PDF, PNG, and JPG document upload
- OCR with Amazon Textract
- Entity recognition with Amazon Comprehend
- Sentiment analysis with Amazon Comprehend
- AI-assisted summarization with Amazon Bedrock
- Event-driven workflow orchestration
Technology Stack
Frontend
Backend
AI Services
Data and Delivery
Infrastructure and Security
Project Status
The original AWS resources and live environment have been decommissioned. The project is presented as an implementation and architecture case study, and the source code remains private.