Research Assistant (SSI)

Interested applicants are invited to apply directly at the NUS Career Portal. Please note your application will only be processed if you apply via NUS Career Portal.

NUS Career Portal link: https://careers.nus.edu.sg/job/Research-Assistant-%28SSI%29/33960-en_GB/

We regret that only shortlisted candidates will be notified.

Job Description

The Research Assistant will develop the core components of the Data Engine Minimum Viable Product (DEM), an end-to-end system for collecting, curating and leveraging large-scale robot manipulation data for training generalist robot policies.

Specific duties include:

• Develop automated data quality assessment methods, including uncertainty quantification and out-of-distribution detection, to identify and filter low-quality or unsafe demonstrations.

• Build and maintain teleoperation and data-collection rigs for real robot platforms, including motion retargeting, inverse kinematics and bimanual/whole-body control.

• Train, fine-tune and benchmark vision-language-action (VLA) and imitation learning policies on curated datasets; run systematic studies on data scale, mixture and augmentation. Optimise policy inference for real-time deployment on physical robots.

• Maintain a well-documented, reproducible research codebase and support dataset and software releases.

• Contribute to publications, technical reports and progress reports to the funding partner; present findings at project and group meetings.

• Mentor student assistants and interns, and collaborate with project partners and other lab members.

Qualifications and Requirements

• Master's degree in Electrical/Electronic Engineering, Computer Science, Mechanical Engineering, Robotics or a related discipline.
• Strong programming ability in Python and C++, with hands-on experience in ROS/ROS2.
• Practical experience with real robot manipulators, including teleoperation, retargeting, inverse kinematics or optimal/adaptive control.
• Experience training deep learning models in PyTorch, with working knowledge of imitation learning / behaviour cloning and vision-language-action or large multimodal models.
• Experience building large-scale data pipelines — dataset curation, standard robot data formats, versioning and quality control — is an advantage.
• Track record of peer-reviewed publications at robotics or machine learning venues (e.g. CoRL, ICRA, IROS, RSS, NeurIPS) preferred.
• Good written and spoken English; able to work independently as well as in a collaborative, multi-disciplinary team.