ISTD PhD Programme
Collaborate with the best minds in a fluid and stimulating setting and conduct breakthrough research that will make a difference to the world.
LOCATION
Singapore University of Technology and Design (SUTD)
PROGRAMME MODE & CANDIDATURE
Full-time (36 – 48 months)
Part-time (36 – 60 months)
Part-time (36 – 60 months)
GRADUATE WITH
Doctor of Philosophy degree under Information Systems Technology and Design (ISTD)
PhD opportunities
| Professor | Title of PhD project | Description | Lab |
| HERREMANS Dorien | Advancing foundational AI models in audio and music | In the Audio, Music, and AI lab (AMAAI), we work on foundational multimodal AI models. We are currently working on new generative audio models, as well as related technologies such as captioning models and transcription models. Some of our previous research includes Mustango, text2midi, video2music, and music2emo. We are looking for students that are very well versed in large multimodal models and proficient in generative models (diffusion, flow-based, autoregressive) to develop state-of-the-art models on topics such as audio mastering and restoration, stem-based generative audio/music/song models, emotionally-informed multimodal models, and joint multimodal embeddings. | Prof Dorien’s website |
| ZHOU Jianying | Securing cyber-physical systems, especially the shipboard OT systems, power grid systems, and water treatment systems | Please refer to https://itrust.sutd.edu.sg/nsoe-destsci-sector-leads/ for more specific research topics carried out in iTrust under the NSoE program: National Satellite of Excellence in Design Science and Technology for Secure Critical Infrastructure. | iTrust |
| LEE Roy | Responsible AI for social good | Please refer to Social AI Studio for the research done by our research studio. We are an interdisciplinary team, and our mission is to foster the design of next-generation socially-enabled artificial intelligence (AI) systems using computational social science, social computing, machine learning, and natural language processing techniques. | Social AI Studio |
| ZHAO Na | Embodied intelligence in the real world | Advance embodied intelligence in AI systems, like robots or autonomous vehicles, by developing robust perception, reasoning, and action components. The focus will be on enabling these systems to understand and interact with dynamic, real-world environments, ensuring reliable operation in open, unstructured settings. | Intelligent Machine Perception Lab |
| SOREMEKUN Ezekiel | Trustworthy software and AI | At the Trustworthy SoftWare & AI (TrustWare) research group, we develop automated methods to evaluate the reliability and trustworthiness of software and AI systems (eg, Code LLMs) using rigorous validation techniques. We assess both functional (eg, correctness) and non-functional (eg, security, robustness, fairness) properties through software testing, debugging, and program analysis. We also build automated tools to support developers and ML engineers in software testing, AI testing, and security analysis. | TrustWare research group |
| SUN Zhu | Trustworthy LLMs for mental healthcare recommendation | An important research topic at the Trustworthy User Modelling and Personalisation Lab (TUMP) at SUTD aims to develop trustworthy large language models (LLMs) that can assist in personalized and responsible mental healthcare recommendations. By integrating large-scale passive sensing data, ethical safeguards, and user-centric design, the system seeks to offer supportive guidance for mental wellness while avoiding harm or bias. Special focus will be placed on explainability, data privacy, and alignment with mental health professionals’ standards, ensuring that the LLMs serve as safe, transparent, and complementary tools in mental health support – particularly in early-stage intervention, resource navigation, and emotional self-care. Please refer to https://sites.google.com/view/zhusun/home for more detailed research topics of our group. | TUMP Lab |
| ZHANG Wenxuan | Inclusive, efficient, and trustworthy large language models | The inclusive NLP (iNLP) Lab at SUTD works on advancing natural language processing and large language models that are inclusive (eg, supporting diverse languages and cultures through multilingual large language models), efficient (eg, making LLMs more accessible through model compression or multi-agent collaboration), while also trustworthy through techniques that improve understanding, safety and robustness of language models. | iNLP Lab |
| CHOO Kenny | Human-AI interaction for health, wellness, and human performance | This research area explores how artificial intelligence can support and enhance human health, wellbeing, and performance through effective interaction design. Projects may span intelligent coaching systems, adaptive human–AI collaboration, digital health tools, or AI-driven insights for physical and cognitive performance. Students will have the opportunity to investigate both the technical foundations of AI systems and the human-centered design principles that ensure trust, accessibility, and impact. The work draws on our lab’s interdisciplinary expertise in human–computer interaction, affective computing, and applied machine learning, with applications ranging from healthcare and fitness to mental wellbeing and workplace performance. | Context-Aware Interaction Lab |
| LI Xiaoli | Advanced time series foundation models | Advanced Time Series Foundation Models use modern machine learning techniques such as self-supervised learning (eg, contrastive learning, masked time-series modelling) to automatically learn useful features from large amounts of unlabelled data and build accurate prediction models. Once trained, these models can be adapted through transfer learning, domain adaptation, and multi-task learning to support different applications, including time series classification, forecasting, equipment condition monitoring, predictive maintenance (PM), and remaining useful life (RUL) prediction. To make the models practical for real-world deployment, methods such as model compression are used to reduce memory and computation needs. The goal of this project is to build a unified model that not only captures rich temporal patterns but can also be efficiently adapted to new tasks and environments, enabling robust and scalable solutions for industrial applications. | Prof Li’s website |
| MASHIMA Daisuke | Digital twins for cyber-physical systems security in smart cities/nation | The project is seeking a PhD student working on the design, implementation, and application of digital twins of a wide range of cyber-physical systems and infrastructure for cybersecurity. The project scope may include, but is not limited to, power grids and EV charging networks, water treatment plants, smart buildings, aviation/airport systems, and biomedical/healthcare systems. The project will explore AI‑based cyber‑physical system modelling to generate high‑fidelity virtual replicas of these systems or infrastructures, which are to be used as a platform for cybersecurity experiments and training, as building blocks of cybersecurity solutions, such as anomaly detection, automated security/resilience assessment through automated red-teaming, and threat intelligence collection and analysis, and as a tool for synthetic data generation for evaluation and benchmarking. | Prof Mashima’s website |
| SIEW Marie | Network-aware optimisation for distributed intelligence in smart cities | Our lab’s research revolves around optimising resources for enabling distributed intelligence in smart cities.
We work at the intersection of networks, wireless communication, machine learning, and mathematical optimisation, designing methods that make edge computing, IoT, and federated learning systems efficient, privacy-preserving, incentive-aware, fair, and sustainable. Topics which students can work on include carbon-aware distributed computing and learning, network optimization for embodied AI multi-agent communication, privacy preserving low-rank fine-tuning of larger models in resource-constrained network settings. |
Prof Siew’s website |
| SONG Peng | Interactive design with mixed reality | The Computer Graphics Laboratory (CGL) at SUTD does research in geometric modelling, computational design, and computational fabrication. We are particularly interested in developing computational methods, algorithms, and tools for modelling and designing fabricable and functional real-world objects as well as optimizing their performance. The research developed in CGL has been applied to a wide variety of applications, including recreation, healthcare, architecture, and engineering. | CGL |