Robotics ML Engineer - Learning & Control

Stealth Robotics; Boston (MA); On-site; USD 120,000 - 190,000; Full time

Benefits offered by this job

Early-stage equity

Publication opportunities

Research collaboration

Job summary

Stealth Robotics is seeking a Member of Technical Staff focused on Learning & Control to advance learned behaviors for physical tasks. You will take ideas from training through hardware experiments, focusing on generalization, adaptation, and recovery when conditions change.

You’ll collaborate with perception, controls, and integration engineers to deploy policies on real robots, design rigorous experiments, and publish when appropriate. Equity is offered for early-stage contributions.

Qualifications

• 2+ years of relevant research or engineering experience, including graduate research.
• MS or PhD preferred; equivalent experience welcome.

Responsibilities

• Develop and evaluate learning-based policies for robotic tasks.
• Improve performance across objects, environments, and task variations.
• Develop recovery behaviors for failed actions and unexpected conditions.
• Design evaluations to measure task success, robustness, and data efficiency.
• Use demonstrations and experimental results to guide data collection and training.
• Collaborate with perception, controls, and integration engineers to deploy policies on hardware.

Skills

Python

Deep learning framework

Learned policies

Real-robot evaluation

Imitation learning

Reinforcement learning

Vision-Language-Action models

Experimental design

Trade-off analysis

Education

MS or PhD in robotics, CS, or related field

Job description

Stealth Robotics is seeking a Member of Technical Staff focused on Learning & Control to advance learned behaviors for physical tasks. You will take ideas from training through hardware experiments, focusing on generalization, adaptation, and recovery when conditions change.

You’ll collaborate with perception, controls, and integration engineers to deploy policies on real robots, design rigorous experiments, and publish when appropriate. Equity is offered for early-stage contributions.