Document Processing Infrastructure for AI Shop on AWS

Company Description and Problem

The AI Shop is an innovative AI solutions provider specializing in advanced Gen AI technologies, and specifically the provision of fine-tuned, self-hosted large language models for privacy-conscious enterprises.


The AI Shop required AWS infrastructure to host their document querying code on AWS and leverage AWS Gen AI capabilities in the most effective and scalable manner.

Proposed Solution

Disraptor proposed a comprehensive AWS solution for the infrastructure that would be required to run The AI Shop’s document processing custom code. The solution leveraged a serverless architecture for scalability and cost-effectiveness, ensuring robustness through asynchronous communication between AWS services.

Key Components of the Proposed Solution

Document Processing Pipeline: Implemented a serverless architecture using AWS Lambda and AWS Textract for document processing.

Data Storage and Management: Stored processed data and validation results in Amazon Aurora DB, ensuring fast and reliable access for analysis and retrieval.

Scalability, Cost Optimization and Performance: AWS’s auto-scaling capabilities were utilized to handle varying loads, ensuring optimal performance and cost optimization during peak periods. The serverless compute model allowed for easy development, deployment, and maintenance of the entire solution. This was achieved by leveraging:

  • AWS ECS: Provisioned ECS cluster so that the custom code developed by The AI Shop can be hosted and seamlessly integrated with the document processing pipeline.
  • API Gateway: Created an API entry point for external access, providing a clean and secure interface for retrieving processed document results and validation outcomes.
  • AWS Bedrock: Provisioned Bedrock for Gen AI capabilities using Anthropic Claude V2, enhancing document validation and retrieval functionalities.
  • Lambda Functions: Implemented multiple Lambda functions to enhance user experience, including feedback mechanisms, prompting, and search functionalities.


The infrastructure was successfully provisioned on AWS by Disraptor, providing The AI Shop with a robust foundation for hosting their custom document querying code. This lays the groundwork for The AI Shop to leverage AWS Gen AI capabilities for enhanced document validation and information retrieval and synthesis. The provisioned infrastructure enables scalability, cost-effectiveness, security and robustness in document processing operations using AWS Lambda and AWS Textract. Overall, this infrastructure sets the stage for The AI Shop to deliver superior document querying and information synthesis solutions to their clients.


Leave a Reply

Your email address will not be published. Required fields are marked *