Executive Summary

The University of Pittsburgh partnered with Quantiphi to modernize its Research Discovery Tool (RDT), migrating it from Azure to AWS. This transformation created a scalable, cloud-native platform that significantly enhances research efficiency and collaboration. The upgraded RDT now features semantic search powered by vector databases and LLMs, allowing researchers to use natural language queries for faster, more accurate discovery. This modern architecture improved user experience, enhanced security, delivered real-time insights and a future-ready foundation for generative AI capabilities.

About the Client

The University of Pittsburgh, partnered with AWS for a Cloud Innovation Center, is a public research institution and a member of the Association of American Universities, recognized for its academic excellence and leadership in research. As part of its commitment to advancing innovation and discovery, the university developed the Research Discovery Tool (RDT) to help researchers efficiently explore and connect with ongoing research across the institution.

Problem Statement

The Research Discovery Tool (RDT) was built to help researchers explore publicly available information and active and completed projects across the institution relevant to their research areas. While the tool was functional, basic design constraints made it challenging for researchers to efficiently locate and analyze relevant research data. These limitations hindered collaboration and slowed discovery across departments. To overcome these challenges and enable a more intuitive, scalable, and high-performing platform, the University sought to modernize the RDT by migrating it to AWS. The goal was to enhance search capabilities, improve the user interface, and lay the groundwork for integrating Generative AI in the future—empowering researchers with deeper insights and streamlined access to institutional knowledge.

Challenges

Limited Scalability: The Research Discovery Tool (RDT) could not accommodate the growing volume of research data and user demand, restricting its ability to scale with the University’s expanding research ecosystem.
Poor Search Accuracy: Researchers faced difficulties in retrieving relevant information efficiently due to limited search functionality and indexing accuracy.
Suboptimal User Experience: The tool’s basic design and complex navigation hindered usability, reducing productivity and engagement among researchers.
Limited Collaboration: The existing system lacked integrated features to promote collaboration, resulting in fragmented research efforts and isolated data access.

Solution

Quantiphi, in partnership with the AWS Generative AI Innovation Center’s (GenAIIC) Partner Innovation Alliance, modernized the University of Pittsburgh’s RDT by migrating it to AWS, transforming it into a platform built for high usability and performance. This effort included a complete re-engineering of the application architecture, a redesigned user interface, and the integration of new features like export functionality and advanced access controls. The upgraded RDT now delivers:

Semantic Search Capability: Implemented semantic search powered by vector databases and large language models (LLMs), enabling researchers to use natural language queries to retrieve contextually relevant results.
Re-architected on AWS: Leveraged a suite of AWS services to deliver faster, more reliable, and highly scalable search performance and platform stability.
Enhanced User Experience: Redesigned the UI for intuitive navigation and improved accessibility, fostering easier exploration of research data.
Robust Access and Export Features: Introduced secure access controls and export capabilities for seamless data sharing and collaboration.

Technologies Used

Amazon ECS

AWS Lambda

Amazon S3

Amazon DynamoDB

Amazon OpenSearch

AWS Glue

Amazon Bedrock

Amazon Cloudwatch

Amazon Cognito

AWS WAF

AWS Secrets Manager

AWS KMS

Results and Business Impact Created

Enhanced Research Efficiency & Accuracy: Researchers now discover relevant projects faster and with greater accuracy via semantic search (powered by vector databases and LLMs)—saving valuable time and improving overall productivity.
Accelerated Insights: The modern AWS architecture provides a highly responsive, contextual search experience, leading to quicker access to meaningful research data and accelerating the entire discovery lifecycle.
Increased Collaboration: Improved accessibility and intuitive navigation foster seamless collaboration among diverse research teams, reducing duplication and promoting knowledge sharing.

As a key Partner Innovation Alliance member of the AWS Generative AI Innovation Center (GenAIIC), Quantiphi leveraged deep expertise in AI and AWS to drive this transformation. This AWS-based solution not only solved immediate scalability and search challenges but also established a secure, scalable foundation for future integration of advanced generative AI capabilities, ensuring the RDT continues to evolve with the University’s dynamic research needs