TwinGPT: AI-Generated Digital Twins with Conversational Interfaces

Successful applicants will join VinUniversity as Master’s students, PhD students, or Research Assistants at the College of Engineering and Computer Science (CECS), VinUniversity.

📍 Project Information

As industrial systems and infrastructure become increasingly complex, the demand for digital twins—virtual replicas used for monitoring, simulation, and optimization—continues to grow. However, current digital twin systems often lack intuitive interfaces and struggle to effectively integrate and reason over heterogeneous data sources.

The TwinGPT project aims to develop next-generation AI-driven digital twins that tightly integrate physical system simulations with large language model (LLM)–based conversational agents, enabling natural and intelligent human–machine interaction.

A key focus of the research is multimodal Retrieval-Augmented Generation (RAG), which fuses diverse data sources including text, real-time sensor streams, time-series data, and simulation outputs. Additionally, an ontology-driven knowledge layer will be developed to formally represent system structures, physical constraints, and operational logic, supporting explainable reasoning, consistency checking, and trustworthy AI behavior.

Students will work on integrating simulation platforms (e.g., Unity), real-time sensor data, and AI models to automatically generate, update, monitor, and optimize digital replicas of real-world systems. Applications span smart manufacturing, robotics, and intelligent infrastructure, positioning this research at the forefront of AI-enabled cyber–physical systems.

Principal Investigators: Prof. Simon Park | Dr. Do Tho Truong 

 📍 Research Objectives

The project focuses on the following key objectives:

(i) Develop digital twin systems integrating physical simulations with large language models

(ii) Design multimodal RAG methods for reasoning over heterogeneous data sources

(iii) Build ontology-driven knowledge representations for explainable and consistent AI reasoning

(iv) Integrate real-time sensor data with simulation platforms for continuous system updates

(v) Develop conversational interfaces for intuitive human–machine interaction

Through these efforts, the project aims to enable intelligent, explainable, and adaptive digital twin systems for complex real-world environments.

📍 Project Contact

For further information, please contact Prof. Simon Park via email [email protected]

Graduate Admissions Contact 

• VinUni Graduate Admissions
•  0978 549 846
• [email protected]

Apply now: https://apply.vinuni.edu.vn/graduate/s/login/ 

#VinUni #VinUniversity#VinUniGraduateResearchExcellenceProgram #VinUniPhD #VinUniMasters#VinUniResearchProjects