Robotics Learning Engineer

Senior Robot Learning Engineer – Large Behaviour Models

Bristol (On-site)

Competitive Salary

Full-time, Permanent

We are working with a cutting-edge robotics company developing advanced humanoid systems for real-world manipulation tasks. They are seeking a Senior Robot Learning Engineer to lead the development of large behaviour models for complex, bi-manual robotic manipulation.

This role sits at the intersection of robot learning, foundation models, and real-world deployment, offering the opportunity to bring state-of-the-art research into production systems.

The Role

You will take ownership of scaling and deploying advanced policy architectures across a humanoid robotics platform, working on:

Large behaviour models (diffusion, transformer-based, VLA/VLM)

Reinforcement learning and imitation learning pipelines

Multi-task, language-conditioned manipulation policies

Sim-to-real transfer and real-world deployment

Key Responsibilities

Design, train, and deploy end-to-end robot learning models

Scale diffusion transformer and VLA-based architectures

Develop generalisable, multi-task manipulation policies

Advance RL pipelines for fine-tuning beyond imitation learning

Build sim-to-real transfer workflows

Collaborate across perception, MLOps, and robotics teams

Contribute to research direction and publish at top-tier venues

Requirements

MSc/PhD in ML, Robotics, Computer Science or similar

Strong experience in robot learning for real-world systems

Expertise in at least two of

Behaviour cloning

Diffusion models

Reinforcement learning

Vision-language-action models

Strong PyTorch and distributed training experience

Proven research or applied impact (publications, systems, OSS)

Nice to Have

Humanoid or bi-manual manipulation experience

Sim-to-real experience (MuJoCo, Isaac Sim, etc.)

Experience with CLIP, DINOv2, or similar models

RL fine-tuning techniques (residual RL, DPPO, etc.)