Staff Reinforcement Learning Engineer

Fruition Group US; California (MO); On-site; USD 150,000 - 210,000; Full time

Job summary

Fruition Group US is seeking a Staff Reinforcement Learning Engineer to lead development of next-generation learning systems for general-purpose robotic manipulation. This role spans simulation, real-world experiments and deployment on physical robots.

You will design RL policies, build scalable training pipelines, mentor engineers, and drive sim-to-real transfer with domain randomization and robust evaluation across tasks, embodiments and environments.

Qualifications

• 7+ years of experience in machine learning, robotics or related fields with hands-on reinforcement learning experience.
• Strong experience deploying RL algorithms for physical robotic systems.
• Deep understanding of reinforcement learning fundamentals, including policy optimisation, value functions, exploration, reward modelling and offline/online learning.
• Experience with robotic simulation environments such as MuJoCo, Isaac Sim or equivalent.

Responsibilities

• Lead the design and development of reinforcement learning policies for complex robotic manipulation and whole-body behaviours.
• Develop scalable training pipelines across simulation and physical robot environments.
• Own RL architecture decisions spanning policy representation, reward design, exploration, training strategy and evaluation.
• Develop approaches combining reinforcement learning with imitation learning, behaviour cloning and foundation-model-based policies.
• Drive sim-to-real transfer, including domain randomization, system identification and robustness to real-world dynamics.
• Develop methods for policy fine-tuning using real-world robot interaction and demonstration data.
• Work closely with perception, controls, data and robotics engineers to integrate learned policies into complete robotic systems.
• Establish rigorous evaluation methodologies for policy performance, robustness, generalization and failure recovery.
• Investigate novel approaches to improving robot learning efficiency and generalization across tasks, environments and embodiments.
• Provide technical leadership across RL initiatives, influencing architecture, research direction and engineering standards.
• Mentor engineers and researchers while remaining deeply hands-on technically.

Skills

Python

C++

PyTorch

JAX

Reinforcement Learning

Robotics

Machine Learning

Policy Optimization

Simulation

Tools

MuJoCo

Isaac Sim

Job description

This is an opportunity to join a highly technical robotics team building general-purpose robotic intelligence. The successful engineer will have significant influence over how reinforcement learning is applied to real-world robots, working at the intersection of machine learning, robotics, simulation and physical AI.

We are seeking a Staff Reinforcement Learning Engineer to lead the development of next-generation learning systems for general-purpose robotic manipulation. This is a highly technical role focused on taking reinforcement learning research from algorithm development and large-scale training through simulation, real-world experimentation and deployment on physical robots.

Responsibilities :

• Lead the design and development of reinforcement learning policies for complex robotic manipulation and whole-body behaviours.
• Develop scalable training pipelines across simulation and physical robot environments.
• Own RL architecture decisions spanning policy representation, reward design, exploration, training strategy and evaluation.
• Develop approaches combining reinforcement learning with imitation learning, behaviour cloning and foundation-model-based policies.
• Drive sim-to-real transfer, including domain randomization, system identification and robustness to real-world dynamics.
• Develop methods for policy fine-tuning using real-world robot interaction and demonstration data.
• Work closely with perception, controls, data and robotics engineers to integrate learned policies into complete robotic systems.
• Establish rigorous evaluation methodologies for policy performance, robustness, generalization and failure recovery.
• Investigate novel approaches to improving robot learning efficiency and generalization across tasks, environments and embodiments.
• Provide technical leadership across RL initiatives, influencing architecture, research direction and engineering standards.
• Mentor engineers and researchers while remaining deeply hands-on technically.

Required Experience

• 7+ years of experience in machine learning, robotics or related fields, including significant hands-on reinforcement learning experience.
• Strong experience developing and deploying RL algorithms for physical robotic systems.
• Deep understanding of reinforcement learning fundamentals, including policy optimisation, value functions, exploration, reward modelling and offline/online learning.
• Strong Python and C++ programming skills, with extensive experience in PyTorch and/or JAX.
• Experience with robotic simulation environments such as MuJoCo, Isaac Sim or equivalent.
• Demonstrated experience transferring learned policies from simulation to real hardware.
• Strong understanding of robot dynamics, control and manipulation.
• Experience working with large-scale robotics datasets and distributed GPU training.
• Ability to take ambiguous research problems and translate them into robust engineering solutions.
• Experience with humanoid or mobile manipulation platforms.
• Experience with whole-body reinforcement learning and contact-rich manipulation.
• Experience combining RL with Diffusion Policies, Vision-Language-Action models or robot foundation models.