IbiDs LLC
Coral Gables, FL
Government Procurement and Contracting Solutions
Amazon S3, AWS Lambda, Amazon Textract, Amazon Rekognition, Amazon SageMaker, Amazon Bedrock, Amazon EC2, Amazon RDS, Amazon CloudWatch, AWS Glue, AWS CloudTrail, AWS Trusted Advisor, AWS WAF
IbiDs LLC, a digital solutions provider for the government procurement and contracting sector, needed to streamline its data cleaning, extraction, and search capabilities. The company faced challenges managing unstructured documents and delivering accurate insights to users. Avahi designed and implemented a GenAI-powered data pipeline and retrieval-augmented generation (RAG) system on AWS. The solution automated PDF and image extraction, delivered structured outputs, and enabled intelligent query responses while ensuring scalability, security, and operational efficiency.
IbiDs LLC is based in Coral Gables, Florida, and provides digital data tools, business intelligence, and process automation tailored to government procurement and contracting. By enabling better insights from unstructured government data, IbiDs supports contractors and agencies in making informed, efficient decisions.
Government procurement involves vast amounts of unstructured documents including PDFs, spreadsheets, and images. IbiDs struggled with transforming this data into structured formats at scale. Without automation, manual extraction and cleaning slowed operations, limited insight accuracy, and risked delays in serving their customers.
In addition, IbiDs wanted to give users an intelligent search experience that could return context-rich answers with citations. Without a robust solution, the company risked falling behind in providing timely, AI-powered intelligence to procurement professionals.
IbiDs selected AWS for its breadth of native AI, ML, and data services along with proven scalability and security. Amazon Textract and Amazon Rekognition offered advanced document and image extraction capabilities, while Amazon Bedrock provided generative AI services and embeddings for building a reliable RAG system. AWS’ global infrastructure and compliance features also aligned with the security requirements of the government sector.
IbiDs chose Avahi, an AWS Premier Tier Partner, for its expertise in architecting GenAI workloads on AWS. Avahi had a proven track record delivering AI and data transformation solutions with rapid deployment. The customer valued Avahi’s ability to integrate advanced document processing, RAG frameworks, and AWS-native services into a seamless production pipeline.
Avahi’s consultative approach included demos, iterative validation, and knowledge transfer, ensuring IbiDs could operate and scale the solution independently post-deployment.
Avahi implemented an end-to-end data cleaning and AI pipeline using AWS services:
– ETL Pipeline: Avahi built an automated ETL process to ingest raw data (PDFs, images, spreadsheets) into Amazon S3. AWS Lambda functions handled data ingestion, pre-processing, and post-processing, with triggers ensuring seamless workflows.
– Data Extraction: Amazon Textract extracted text and tables from documents, while Amazon Rekognition and SageMaker image recognition improved quality and accuracy. Outputs were transformed into structured JSON and stored in S3.
– Entity Recognition and RAG: Avahi embedded structured outputs into a vector database using Amazon Bedrock embeddings. Lambda jobs indexed and retrieved embeddings, enabling Bedrock-powered RAG to generate precise answers with citations.
– APIs and Integration: A standalone API deployed on Amazon EC2 provided endpoints for integrating AI-powered responses directly into IbiDs’ existing workflows.
– Monitoring and Security: Avahi configured Amazon CloudWatch for logging and monitoring, and enforced AWS best practices for VPC, IAM, and security groups.
– Knowledge Transfer: Comprehensive documentation and training sessions prepared IbiDs’ team to manage the system.
– ETL pipeline to transform unstructured inputs into structured data
– Enhanced smart assistant with RAG capabilities
– AI-based image recognition system for entity recognition
– API endpoints for seamless integration with IbiDs’ environment
– Secure, scalable infrastructure following AWS standards
– Documentation and knowledge transfer sessions
The solution enabled IbiDs to automate and scale document processing, reducing manual work and increasing accuracy. Users can now query government procurement data with natural language and receive context-backed answers. This improved both internal efficiency and customer satisfaction, positioning IbiDs as a technology leader in its sector.
Metrics
– Automated ingestion and extraction of PDFs, spreadsheets, and images into structured JSON
– End-to-end ETL pipeline and AI workflow completed within 8–10 weeks
– Accuracy and scalability validated against AWS Bedrock model benchmarks
Founder, Bravo Foxtrot
