Robot Learning Engineer: Deploy Policies to Real Robots

Dexmate; Singapore; On-site; SGD 120,000 - 180,000; Full time

Job summary

Dexmate is building the next layer of robot intelligence, offering a platform that streamlines learning-based control for real-world robots. This role focuses on designing and deploying learned policies across the full robot learning pipeline, from data ingestion to hardware deployment.

The ideal candidate has a strong background in robotics, machine learning, and a track record of moving research into deployed systems.

Qualifications

• Advanced degree in Robotics, ML, CS or equivalent practical experience.
• Hands-on ML model training and deployment, preferably with PyTorch.
• Background in robot learning, imitation or reinforcement learning.
• Experience taking models from research to deployment in real systems.

Responsibilities

• Design, train, and evaluate learned policies for manipulation and whole-body control.
• Deploy and validate models on physical robots, tuning for real-world performance.
• Iterate between simulation and real robots, collecting data and debugging failures.
• Own infrastructure for model training: scheduling, checkpoints, metrics, logging.
• Build tooling for debugging, visualization, and experiment analysis.
• Collaborate with controls, embedded, and hardware teams to deploy reliable systems.

Skills

Advanced degree

ML model training

Robot learning

Ownership mindset

Research-to-deployment

Education

Master's or PhD in Robotics/ML/CS

Tools

PyTorch

Protobuf/MCAP

Job description

Dexmate is building the next layer of robot intelligence, offering a platform that streamlines learning-based control for real-world robots. This role focuses on designing and deploying learned policies across the full robot learning pipeline, from data ingestion to hardware deployment.

The ideal candidate has a strong background in robotics, machine learning, and a track record of moving research into deployed systems.