Scientist (SRA/ARM), SIMTech

A*STAR - Agency for Science, Technology and Research; Singapore; On-site; SGD 110,000 - 170,000; Full time

Job summary

A*STAR - Agency for Science, Technology and Research in Singapore invites applications for an Embodied AI Scientist to advance fundamental and applied research in humanoid robotics. The role emphasizes actuator-centric embodied intelligence and learning-enabled control closely tied to physical dynamics, sensing and actuation.

The candidate will work in a multidisciplinary group spanning robotics, control, artificial intelligence and mechatronics, developing novel control strategies and

Qualifications

• PhD in mechanical, electrical, control, or mechatronics engineering.
• 3–5 years of relevant work experience preferred; fresh graduates encouraged.
• Strong foundation in robotics, AI, and control systems.

Responsibilities

• Research embodied AI methods for humanoid robots, focusing on actuation, compliance and energy-efficient motion.
• Develop learning- and model-based control strategies accounting for actuator dynamics.
• Integrate multi-modal sensing with learning control architectures (reinforcement learning).
• Address controller tuning transfer via system identification and adaptive control.
• Investigate safe and constrained learning for locomotion and manipulation under uncertainty.
• Integrate mechanical transmission, actuators and control into humanoid robots to demonstrate embodied intelligence.
• Design experiments using humanoid joint modules, actuation testbeds and full-body platforms.

Skills

Humanoid robotics

Reinforcement learning

Actuator dynamics

Model-based control

Embedded control

Python

C/C++

MATLAB

Education

PhD in Mechanical, Electrical, Control or Mechatronics Engineering

Tools

Ansys Maxwell

Comsol Multiphysics

JMAG

Job description

About The Role

We are seeking for an Embodied AI Scientist to conduct fundamental and applied research in humanoid robotics, with a particular emphasis on actuator-centric embodied intelligence. The successful candidate will contribute to the development of learning-enabled control and decision-making frameworks that are deeply integrated with the physical dynamics, sensing and actuation of humanoid robotic systems. You are expected to work in a multidisciplinary research group spanning robotics, control, artificial intelligence and mechatronic system

Role Summary

We are seeking for an Embodied AI Scientist to conduct fundamental and applied research in humanoid robotics, with a particular emphasis on actuator-centric embodied intelligence. The successful candidate will contribute to the development of learning-enabled control and decision-making frameworks that are deeply integrated with the physical dynamics, sensing and actuation of humanoid robotic systems. You are expected to work in a multidisciplinary research group spanning robotics, control, artificial intelligence and mechatronic system

Job Description

• Conduct research on embodied AI methods for humanoid robots, focusing on joint-level actuation, compliance, force interaction and energy-efficient motion
• Develop learning-based and model-based control strategies that explicitly account for actuator dynamics, nonlinearities and other physical constraints
• Integrate multi-modal sensing which includes closed-loop control of position, force/torque, tactile and visual feedback as well as deployment of learning control architectures such as reinforcement learning
• Adapt Controller tuning transfer challenges through system identification, domain adaptation and adaptive control methods
• Investigate safe and constrained learning approaches for humanoid locomotion and manipulation under uncertainty.
• Integration of the mechanical transmission, actuator and control system into the humanoid robot to demonstrate the embodied intelligence
• Design and conduct experimental studies using humanoid joint modules, actuation testbeds and full-body robotic platforms

JOB REQUIREMENT

• Ph.D. degree in Mechanical, Electrical, Control or Mechatronics Engineering
• 3-5 years of work experience or similar is preferred
• Fresh graduates are also encouraged to apply

We Value The Following Skills And Experience

• Experience with the development and control of humanoid robotics systems
• Knowledge of machine learning applications in motion control, such as reinforcement learning will be a plus
• Good understanding of the various actuation principles and up to date with state-of-the-art actuation technologies
• Experience with the design and integration of BLDC and/or PMSM electric motors in robotics joint module
• Familiarity with numerical simulation software tools e.g. Ansys maxwell, Comsol Multiphysics, JMAG
• Experience with the prototyping and assembly of actuator components e.g. frameless motor assembly
• Good working knowledge of digital control techniques and hardware implementation in embedded controllers
• Proficiency in programming languages such as C, C++, Python, or MATLAB.
• Familiarity with industrial communication protocols (e.g., CAN, EtherCAT, Modbus).
• Excellent problem-solving and analytical skills
• Good communications skills and the ability to work in a team