Germany - LLM in-site search

Description

Introduction
The European Centre of Excellence for Civilian Crisis Management e.V (European Centre of Excellence for Civilian Crisis Management (coe-civ.eu) is seeking a vendor to develop and implement an AI-powered search engine for its Knowledge Hub. The search should be powered by a large language model (LLM) that provides a semantic search experience for its users. This system should improve user search capabilities by offering answers based on natural language processing (NLP) while handling multiple sources of data, misspellings, and user-specific content. The AI-powered solution should be secure, user-friendly, and compliant with EU data privacy standards.
Project Overview
The goal of this project is to integrate an AI-driven search tool on our Knowledge Hub, an online library aimed at gathering all relevant documents on the topic of civilian crisis management. Each document is hosted on a page with a summary of the document, topical tags and either a link to the original source (directing to an external webpage), or a link to a locally hosted document, when we hold the copyrights. The tool should enable users to ask a question and get an answer based on all documents it has access to on the Hub. The tool should cater to two types of users (logged-in and anonymous).
These terms of reference outline the technical requirements, data security, administrative needs, and pricing expectations for the project.
Required features

Editor Independence
Website editors should be able to manage and update the system independently, without relying on external support for regular operations.
Post-Implementation Support
The selected vendor should offer post-implementation support, including troubleshooting, updates, and potential system enhancements.
Collaboration with Web Agency
The vendor should be willing to collaborate with our web development agency, which is currently managing our website development and infrastructure.

Website Platform
Our website is based on Typo3 (PHP), and the search tool must seamlessly integrate with this CMS.
Optional features

User Analytics
The system should track and provide insights into user behavior, search trends, and system performance.
Vendor Requirements
Vendors responding to this tender should provide the following information:
Full Terms of Reference and instructions to bidders are available on our website.

Opportunity closing date

12 November 2024

Value of contract

to be confirmed