Research Fellow

About SUTD

SUTD is the world's first Design.AI university. With Design.AI, artificial intelligence is treated as a partner and a member of the team - not just a tool. As a result of this unique SUTD treatment, AI and humans brainstorm, spar and prototype together, resulting in solutions that are elevated several-fold. This human-AI team concept has been made possible because of SUTD's unique cohort-based interdisciplinary pedagogy. As a trailblazer in the field of design and technology education and research, SUTD has been pioneering innovative programmes and initiatives since its formation 16 years ago. As part its new mid-term growth strategy called SUTD Leap, it will work closely with industry to co-create solutions in both education and research - for a better world.

About the Job

We are seeking a highly motivated Research Fellow to join our team on the project Reliable and Efficient AI under Distribution Shift. The successful candidate will work on developing training-free methods to improve the reliability, robustness, and efficiency of modern AI systems, including deep learning models, time series foundation models, and industrial AI applications.This role offers an opportunity to contribute to cutting-edge research in out-of-distribution generalization, model calibration, and trustworthy AI, with applications in manufacturing and real-world data systems.

What You'll Do

• Conduct research on model reliability and robustness under distribution shift, including OOD detection and uncertainty estimation
• Develop post-hoc and training-free calibration methods leveraging model internals (e.g., activations, gradients, spectral properties)
• Design and implement algorithms for bias correction and efficient model adaptation
• Apply methods to time series data and industrial fault diagnosis (e.g., predictive maintenance, RUL)
• Evaluate approaches on real-world and benchmark datasets
• Publish in top-tier conferences/journals (e.g., NeurIPS, ICCV, IJCAI)
• Collaborate with interdisciplinary teams and industry partners

Who We're Looking For

• PhD in Computer Science, Electrical Engineering, or related field
• Strong background in machine learning, deep learning, or domain adaptation
• Experience in OOD detection, uncertainty estimation, or model robustness
• Proficiency in Python and deep learning frameworks (e.g., PyTorch)
• Solid mathematical foundation (e.g., optimization, statistics, linear algebra)
• Proven publication record in top-tier venues
• Ability to work independently and collaboratively in research projects