Principal Robotics AI Engineer (Loco-Manipulation & Whole-Body Control)

Job Overview

AI.DA STC (Strategic Technology Centre)'s NEAR (Next-gen Edge AI & Robotics) Lab is building a deep ground robotics research capability. We're looking for a Principal Engineer to anchor our loco-manipulation and whole-body control work. This involves making legged platforms reliable for contact-rich manipulation and cooperative tasks in unstructured environments.

This is an engineering-first role with active research participation. The primary measure of impact is shipping working systems on real hardware. Publications happen alongside that work, co-authored with academic collaborators and lab engineers when the technical contribution warrants it. You'll join a small ground robotics team, partner with our embodied AI lead, and work alongside academic collaborators, startups and robotics vendors in Singapore and abroad.

This role reports initially to the Lab Director. As the ground robotics team grows over the next 12 months, the reporting line will evolve; we'll be transparent about this trajectory throughout the hiring process.

Loco-Manipulation & Whole-Body Control

• Whole-Body Control & RL for Legged Systems: Design and ship the WBC stack for our legged platforms; coupled base-arm controllers, contact-aware control, and hybrid policies combining RL with model-based control. Sim-to-real for contact-rich tasks, training in Isaac Sim or MuJoCo, deployment onto edge compute.
• Contact-Rich Manipulation: Build manipulation policies for tasks where balance, contact force, and locomotion are coupled: door opening, valve turning, payload transport, debris handling. Force and impedance control on real hardware.
• Cross-Embodiment Transfer: Lead the lab's work transferring policies across embodiments: bimanual demonstrations to single-arm quadrupeds, quadruped manipulation to humanoid.
• Technical Direction & Mentorship: Set the technical direction for the ground robotics manipulation stack and develop more junior engineers through code review and direct technical mentorship.

Wider Scope

• Locomotion & Autonomous Navigation: Partner with the embodied AI lead on rugged-terrain locomotion and on integrating loco-manipulation with the autonomous navigation stack.
• Multi-Robot & Cooperative Manipulation: Contribute to multi-robot manipulation as the lab scales from single-arm to distributed bimanual systems and cooperative payload transport, including the deployment infrastructure (ROS 2 multi-robot architecture, RMF, mesh networking) that makes multi-robot field trials work.

Required Qualifications

• Education & Experience: PhD strongly preferred in Robotics, Mechanical Engineering, Computer Science, Electrical Engineering, or related field; or Master's with strong industry-research track record. 8+ years of post-graduate experience in legged robotics, mobile manipulation, or whole-body control. Shipped systems and field deployment weighted as heavily as academic credentials.
• Whole-Body Control & RL Depth: Direct implementation experience with WBC (classical MPC/QP, learned, or hybrid) and hands-on deployment of RL policies onto real legged platforms, with demonstrated ability to close the sim-to-real gap on hardware.
• Contact-Rich Manipulation: Hands-on experience with force, impedance, or admittance control on real arms, ideally on mobile bases.
• Real-Robot Track Record: Multi-year pattern of taking engineering problems to working solutions on hardware: bring-up, commissioning, field debugging, on-robot diagnostics. Demonstrated ability to adapt open-source robotics repositories to real platforms and make them survive field conditions.
• ROS 2 & Multi-Robot Systems: Strong working knowledge of ROS 2 including multi-robot architecture, DDS configuration, and real-world debugging of distributed systems. Familiarity with Open-RMF or equivalent. Direct experience deploying multi-robot systems on real hardware.
• Software: Strong C++17/20 and Python in Linux environments. Fluent with modern AI coding assistants and clear-headed about when to use them.

Preferred Qualifications

• Industry Research Experience: Prior work at industrial research labs that ship to hardware, e.g. NVIDIA Robotics, Toyota Research Institute, or equivalents.
• Advanced Manipulation or humanoid: Hands-on work with industrial/research arms (Franka, UR5e, uFactory), or with humanoid platforms for contact-rich manipulation.
• Collective & Cooperative Manipulation: Prior work on multi-robot cooperative manipulation, collective transport, or coordinated loco-manipulation across multiple legged platforms.
• World Models & Learned Dynamics: Exposure to world models or learned dynamics models for manipulation planning and multi-robot coordination.
• VLA, Foundation Models & Imitation Learning: Experience deploying VLAs on real robots; familiarity with bimanual imitation-learning pipelines (Mobile ALOHA or equivalent).
• Tactile Sensing & Simulation Depth: Prior work with tactile sensors, F/T feedback, or compliant grippers; depth in Isaac Sim, MuJoCo, or Isaac Lab including domain randomisation.
• Research Output: Peer-reviewed publication at ICRA, CoRL, RSS, RA-L, or IROS.
• Field Robotics: Context in safety-critical, defence-relevant, or extended field-deployed robotics.