Reinforcement Learning & Controls Research Scientist- Spot Behavior

Boston Dynamics; Waltham (MA); On-site; USD 177,000 - 225,000; Full time

Benefits offered by this job

Medical benefits

401(k)

Paid time off

Annual bonus structure

Job summary

Boston Dynamics is seeking an engineer to develop reinforcement learning solutions for our quadruped robots. This role involves designing, training, and deploying RL policies that enhance mobility and robustness in challenging terrains.

The ideal candidate will hold an MS or PhD in a relevant field and have 3–6 years of experience in deploying RL control policies. Proficiency in Python and C++ is required, along with a solid foundation in classical control theory. A generous benefits package is offered.

Qualifications

• 3–6 years of experience deploying RL control policies on physical hardware.
• Strong foundations in classical control theory and stability analysis.
• Familiarity with real-time software constraints and control loop design.

Responsibilities

• Design and deploy RL systems to improve Spot's mobility.
• Tune low-level controllers interfacing with learned policies.
• Build and maintain simulation environments for training policies.

Skills

Reinforcement Learning

Control Theory

Python

C++

Deep RL Frameworks

Education

MS or PhD in Robotics, Mechanical Engineering, Computer Science

Tools

Isaac Sim

MuJoCo

Job description

At Boston Dynamics, we are pushing the boundaries of what legged robots can do in the real world. The Spot Behavior team is building next-generation locomotion and mobility capabilities, and we are seeking a curious, driven engineer to develop reinforcement learning solutions that run directly on our quadruped platforms. In this role, you will design, train, and deploy RL policies that integrate tightly with Spot's control stack to deliver robust, agile behavior across real-world environments.

Day-to-Day Activities

• Design and deploy RL systems that improve Spot's mobility and robustness across challenging terrain.
• Tune and validate low-level controllers (e.g., PD/PID, whole-body control) at the interface with learned policies.
• Build and maintain simulation environments (e.g., Isaac Sim, MuJoCo) to train and validate policies before hardware deployment.
• Analyze robot data logs to diagnose control failures and iterate on controller design.
• Test and debug directly on our in-house Spot fleet, taking a first‑principles approach to failure analysis.
• Write production‑ready code in Python and C++.

We Are Looking For

• MS or PhD in Robotics, Mechanical Engineering, Computer Science, or related field.
• 3–6 years of experience deploying RL or learning‑based control policies on physical hardware.
• Strong foundations in classical control theory, including stability analysis, state estimation, and low‑level actuation.
• Experience with real‑time software constraints and control loop design.
• Proficiency in Python and C++.
• Familiarity with modern deep RL frameworks (e.g., PyTorch, RLlib).

Nice To Have

• Experience with legged robotics or contact‑rich locomotion systems.
• PhD in a relevant field.
• Familiarity with whole‑body control or model predictive control (MPC) for legged systems.
• Experience with state estimators (e.g., EKF, contact estimation) in robotics.
• Experience with sim‑to‑real transfer and domain randomization.

The base pay range for this position is between $177,000 to $225,000 annually. Base pay will depend on multiple individualized factors including, but not limited to internal equity, job related knowledge, skills and experience. This range represents a good faith estimate of compensation at the time of posting. Boston Dynamics offers a generous Benefits package including medical, dental vision, 401(k), paid time off and an annual bonus structure. Additional details regarding these benefit plans will be provided if an employee receives an offer for employment.