Domain-Grounded Conversational AI Methodology for Technical Document Understanding: curated-corpus retrieval-augmented generation, tool-invoking agent workflow and expert-validated evaluation protocol for specialised large language models

Licensing negotiation in progress

Published Date: 5 November 2026

Closing Date: 19 November 2026

We are pleased to invite expressions of interest for the commercializing of the Domain-Grounded Conversational AI Methodology for Technical Document Understanding: curated-corpus retrieval-augmented generation, tool-invoking agent workflow and expert-validated evaluation protocol for specialised large language models developed by Prof. SONG Jun of Assistant Professor of the Academy of Geography, Sociology and International Studies of HKBU. (KTO ref: KH-DGCA)

Description

This technology is a validated framework for adapting general-purpose large language models to specialized technical or scientific domains without costly retraining, while strictly controlling hallucinations and ensuring every output is traceable to an authoritative source. It is domain-agnostic, relying on the document corpus, analytical tools, and benchmark questions as the only domain-specific inputs.

Note

This invitation of expression of interest is made without prejudice.  This invitation does not constitute a tender, and HKBU is not bound to accept any offer, regardless of monetary value. The final decision regarding the commercialization of the technology rests with HKBU.

For inquiries, please contact:
Ms Vivian Kwan (email: vivian_stkwan@hkbu.edu.hk)
Knowledge Transfer Office