The Hyundai-NTU-A*STAR Corp Lab is seeking a Research Fellow to lead research on an LLM-based smart assistant system for automotive equipment maintenance. This project aims to develop technologies for multi-query strategies and the fusion of information from diverse sources, including equipment manuals, forums, databases, and knowledge graphs. The role involves leading a team of researchers to propose novel ideas and develop innovative solutions to complex technical problems. Key Responsibilities: Lead research on LLM-based smart assistant systems; manage a team of researchers; conduct literature reviews on AI and machine learning for decision-making; design AI solutions and search modules for information fusion; evaluate system performance against existing models; and author research papers to report findings. Job Requirements: A PhD in computer science or a related field is preferred. Candidates must possess expertise in computer programming, knowledge of machine learning and data science, and a proven research track record evidenced by publications, conference papers, or patents. Strong analytical, communication, and teamwork skills are essential, along with the ability to design experimental plans and initiate multidisciplinary collaborations. The role requires the ability to oversee project progress and a willingness to work flexible hours.
For full details about this role, click “Apply on company site” to view the employer’s original posting.
Skills for this role
Large Language ModelsMachine LearningData ScienceComputer ProgrammingInformation FusionKnowledge GraphsResearchTechnical Writing
Your skill match
Checking your profile…
YOUR NEXT STEP
Get interview-ready for this role
A focused preparation guide, built around this job’s responsibilities and requirements.
✦ AI-generated guide
Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.
01
Understand the role
What matters most
Architecting Multi-Source Information Fusion
The core technical priority is designing a search module capable of synthesizing data from heterogeneous sources including equipment manuals, internet forums, databases, and knowledge graphs. You must demonstrate how to use LLMs to effectively merge and prioritize these disparate streams for automotive maintenance applications.
Technical Leadership and Research Direction
As a Research Fellow, you are expected to lead a team of researchers in proposing novel ideas and developing innovative solutions. This requires balancing high-level project oversight with the ability to guide the team through complex technical problem-solving and experimental design.
Rigorous Performance Evaluation
The role requires evaluating and comparing proposed systems against existing models. You must prioritize the development of robust metrics to measure the effectiveness of the LLM-based assistant in real-world maintenance scenarios and document these findings in high-quality research papers.
Likely challenges to think through
Managing Data Heterogeneity and Noise
Integrating structured data like knowledge graphs with unstructured, potentially noisy data from internet forums presents a significant challenge. Prepare by discussing strategies for data cleaning, entity alignment, and ensuring the LLM maintains high accuracy when dealing with conflicting information across sources.
Bridging Academic Research and Industrial Application
The project is part of a corporate lab (Hyundai-NTU-A*STAR), which implies a need for research that is both novel and industry-ready. You may face difficulty in ensuring your theoretical models are scalable and practical for automotive maintenance environments; prepare by emphasizing your experience in translating experimental methods into functional, deployable solutions.
03
Practice your answers
Technical practice
LLM-based Information Fusion Strategies
What to coverDiscuss how you would architect a system to fuse information from heterogeneous sources like equipment manuals, forums, and knowledge graphs. Address the trade-offs between RAG (Retrieval-Augmented Generation) architectures and fine-tuning for domain-specific maintenance tasks.
Multi-Query Strategy Design
What to coverExplain your approach to designing a multi-query strategy for an LLM-based assistant. Consider how to handle query decomposition, query expansion, and the subsequent merging of results to ensure high precision in technical maintenance contexts.
Knowledge Graph Integration with LLMs
What to coverDescribe how you would leverage knowledge graphs to ground LLM outputs in an automotive maintenance domain. Discuss methods for mapping unstructured data to graph structures and using these graphs to improve the reasoning capabilities of the assistant.
System Performance Evaluation Metrics
What to coverHow would you design an evaluation framework to compare your proposed system against existing models? Suggest specific metrics for measuring the effectiveness of information prioritization and the accuracy of maintenance recommendations.
Experimental Design for AI Research
What to coverDetail your process for formulating a research hypothesis and designing an experimental plan for a novel AI solution. Emphasize how you ensure reproducibility and statistical significance in your findings.
Handling Noisy Data Sources
What to coverMaintenance data often comes from unstructured or noisy sources like internet forums. Explain how you would implement pre-processing or filtering modules to ensure the LLM receives high-quality context for its responses.
Behavioural practice
Leading Research Teams
What to coverDescribe a situation where you led a team of researchers to solve a complex technical problem. Focus on how you fostered innovation, managed disagreements on technical direction, and ensured the project met its research objectives.
Managing Multidisciplinary Collaborations
What to coverProvide an example of a time you initiated or managed a collaboration across different disciplines. Explain how you aligned the goals of different stakeholders and integrated diverse expertise to achieve a successful research outcome.
Adapting to Novel Research Challenges
What to coverTell me about a time when your initial research approach failed or hit a significant roadblock. How did you adapt your strategy, incorporate new ideas, and eventually overcome the challenge?
Communicating Complex Technical Findings
What to coverDescribe a time you had to explain complex AI research findings to a non-expert audience or stakeholders. How did you ensure your communication was effective while maintaining technical accuracy?
02
Sharpen your knowledge
Topics to prepare
Information Fusion and Retrieval Augmented Generation (RAG)
Review architectures for integrating heterogeneous data sources like knowledge graphs, databases, and unstructured manuals. Exercise: Design a conceptual framework for a multi-query system that prioritizes information from conflicting sources, as this is central to the project's requirement for an LLM-based smart assistant.
Experimental Design for AI Systems
Study methodologies for benchmarking LLM performance in domain-specific tasks. Exercise: Draft an experimental plan to evaluate the effectiveness of an LLM-based search module against a baseline, focusing on metrics for information accuracy and prioritization in maintenance contexts.
Skills to revise
Knowledge Graph Integration
Practice linking unstructured text data with structured knowledge graphs. Exercise: Create a small-scale prototype that uses a knowledge graph to ground LLM responses, ensuring the assistant provides accurate technical maintenance advice.
Technical Research Communication
Refine the ability to synthesize complex AI research into clear, actionable findings. Exercise: Summarize a recent paper on multi-source information fusion, focusing on how the findings could be applied to industrial equipment maintenance.
Advanced Programming for AI
Demonstrate proficiency in implementing AI pipelines. Exercise: Develop a Python script that orchestrates a multi-query search across a mock database and a knowledge graph, demonstrating the ability to handle diverse data streams.
04
Your preparation checklist
Synthesize a research proposal for multi-source information fusion
Prepare a brief technical outline demonstrating how you would architect an LLM-based system to fuse data from manuals, forums, databases, and knowledge graphs for automotive maintenance, specifically addressing multi-query strategies.
Curate your portfolio of research outputs
Organize your publications, conference papers, and patents to highlight your expertise in machine learning, data science, and AI, specifically focusing on projects that involved experimental design and hypothesis testing.
Develop a strategy for team leadership and project oversight
Prepare to discuss your experience or approach to managing a team of researchers, including how you track project progress, mentor junior members, and foster multidisciplinary collaboration in a research-intensive environment.
Formulate questions regarding the Hyundai-NTU-A*STAR collaboration
Prepare insightful questions about the specific technical challenges of the automotive maintenance domain, the expected performance metrics for the smart assistant, and how the team balances academic research with industry-driven goals.
Review literature on LLM-based decision-making systems
Conduct a targeted literature review on current state-of-the-art methods for LLM-based information retrieval and fusion to demonstrate your ability to stay current with rapid advancements in AI.
Availability last checked: 18 Oct 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy
Share this opportunityKnow someone who would be a great fit?