Job Description

The Safety and Resilience Research Unit (SaRRU) at the NUS College of Design and Engineering invites applications for the position of Research Fellow (LLM for Design-for-Safety). The successful candidate will join the research project “Multimodal LLM-driven Intelligent Design-for-Safety Advisory System (MIDAS),” which aims to develop an innovative AI-powered advisory system to enhance Design-for-Safety (DfS) practices in Singapore.

The Research Fellow will be responsible for, and work closely with, the Principal Investigator and the project’s existing Research Fellows to ensure the successful completion of the project. Based on the project's needs for extensive data handling, model development, and stakeholder engagement, the role will involve significant responsibilities in development of the knowledge management and reasoning core of MIDAS as well as project management support.

Key responsibilities include:
•    Assisting with the collection, curation, and processing of multimodal data, and integrating multimodal inputs (text, images, and Industry Foundation Classes (IFC) data) into the MIDAS operational workflow.
•    Supporting the development of a national DfS knowledge for retrieval and reasoning in Large Language Models (LLMs).  
•    Developing, evaluating, and refining Retrieval-Augmented Generation (RAG) architectures that ground LLM outputs in DfS domain knowledge and project data.
•    Fine-tuning and adapting LLMs for the DfS domain and applying model-agnostic techniques to improve generalisation across tasks, project types, and data conditions.
•    Providing project management and administrative support, such as coordinating with stakeholders and preparing project documentation, meeting minutes as well as progress reports.
•    Contributing to the dissemination of research findings through reports and academic publications.
•    Carrying out other tasks required by the PI.

Job Requirements

•    A PhD degree in a relevant discipline such as Civil Engineering, Construction or Safety Management, Computer Science, Information Systems, or a related field from a reputable university.
•    Proficiency in programming for multimodal data analysis and LLMs (e.g., building RAG architectures, fine-tuning models, etc.).
•    Familiarity with construction safety, Design for Safety (DfS), and knowledge management is highly desirable.
•    A strong publication record in top academic journals and strong written and oral communication skills in English.
•    Excellent organizational, interpersonal, and communication skills, with the ability to coordinate effectively with various stakeholders.