LINEARK Hiring! Full Time Robotics Engineer in Federal Territory - Ricebowl

Job Summary

Own the physical layer across every humanoid robot LINEARK operates. Bring up new humanoid platforms, map and reverse engineer manufacturer control interfaces, own the motion data path from robot to processed dataset, and act as the technical reference point for hardware behaviour across our teleoperation, computer vision, motion data pipeline and video-to-motion development tracks.

Key Responsibilities

- Bring up and integrate new humanoid robot platforms; map and document their control interfaces and behaviour

- ⁠Reverse engineer undocumented manufacturer stacks across multiple robot manufacturers

- ⁠Own the motion data pipeline end to end: capture, ingestion, processing, quality assurance

- ⁠Diagnose, evaluate, repair, and replace robot hardware

- ⁠Provide hardware-level technical support on teleoperation, computer vision, and video-to-motion systems

- ⁠Build and maintain internal technical documentation and knowledge base

- ⁠Support technical setup, sandbox environments, and operations at activations

Requirements

- Relevant experience in robotics, mechatronics, or embedded systems

- ⁠Direct hands-on experience with at least one commercial humanoid or legged robot platform. Simulation-only experience will not be sufficient

- ⁠ROS 2 at working depth, including practical debugging of DDS QoS, discovery, or transport issues in a live system

- ⁠Demonstrated ability to work with thin or absent manufacturer documentation: reading source, analysing motion traffic, or tracing protocols to determine undocumented behaviour

- ⁠Strong Python and functional C++; Linux native

- ⁠Solid understanding of robot motion data: joint state, IMU, kinematics, coordinate frames and transforms, sampling rates and clock drift

- ⁠Hardware competence: able to open, diagnose, evaluate, and replace robot components

- Based in Kuala Lumpur; on-site presence required for hardware work

Preferred Experiences

- Motion capture and retargeting formats: BVH, LAFAN1, MuJoCo, Isaac Sim

- ⁠Real-time control systems, low-latency networking, teleoperation

- ⁠High-rate time-series data pipelines: Parquet, Arrow, Polars

- ⁠Motor and actuator tuning, gait and balance controller development

- ⁠Experience across two or more robot manufacturers' stacks