If your idea of robotics research is publishing a paper and moving on to the next experiment, this probably isn't for you.

This is a Founding Robotics ML Research Engineer opportunity for someone who wants to take the latest advances in robot learning and turn them into systems that work reliably on physical robots.

The opportunity

You'll join an early-stage robotics company developing autonomous systems for demanding real-world environments.

As a Founding Robotics ML Research Engineer, you'll sit at the intersection of robotics research and production engineering. You'll develop and deploy learning-based approaches for manipulation and autonomous behaviour, while building the infrastructure required to collect data, train models and run them reliably on physical systems.

The role is particularly suited to someone who enjoys the complete loop: identifying a promising research direction, experimenting with new approaches, training models, deploying them to hardware and then iterating based on what happens in the real world.

You'll work with modern learning approaches including behaviour cloning, reinforcement learning, diffusion-based policies and vision-language-action models, while grounding those techniques in classical robotics principles such as manipulation, control and motion planning.

Responsibilities

  • Develop and deploy learning-based policies for complex robotic manipulation tasks
  • Research and apply modern approaches including behaviour cloning, reinforcement learning, diffusion policies and vision-language-action models
  • Build and lead the software infrastructure supporting robotic data collection, model training and real-time inference
  • Translate emerging robotics learning research into reliable, production-ready systems
  • Experiment with new architectures, training strategies and data-generation approaches
  • Evaluate model performance on physical robots and iterate based on real-world results
  • Work closely with robotics engineers across perception, controls, planning and software
  • Develop systems capable of running learning-based models reliably on real robotic hardware
  • Contribute to the technical direction of the company's robotics learning stack
  • Help bridge the gap between research prototypes and robust deployed systems

Why this role

This is a Founding Robotics ML Research Engineer role for someone who wants to work at the frontier of robotics learning without being separated from the physical systems that ultimately have to use it.

You'll have the freedom to explore new research directions, but the end goal isn't simply a benchmark result. Your work will need to survive contact with real hardware.

That creates a particularly interesting feedback loop: research ? training ? deployment ? failure analysis ? iteration.

You'll also have substantial influence over the infrastructure and technical foundations supporting the company's ML and robotics systems. As an early engineering hire, your decisions will help shape how the organisation approaches data, experimentation, training and deployment.

For someone who wants to see their research become part of a working robotic system, rather than sitting indefinitely in a research environment, this is a chance to have a significant impact.

What we're looking for

We're looking for a Founding Robotics ML Research Engineer with strong experience in machine learning for robotics and a genuine interest in deploying learning-based systems on physical robots.

You'll ideally have:

  • A background in Robotics, Computer Science, Machine Learning or a related technical discipline
  • An MS or PhD in a relevant field is preferred
  • Experience training large-scale manipulation models
  • Practical experience with approaches such as diffusion policies, vision-language-action models or other modern robot learning architectures
  • Experience developing and deploying learning-based control or perception algorithms on real robotic systems
  • Strong knowledge of modern ML tools, architectures and development workflows
  • A solid understanding of classical manipulation, control theory and motion planning
  • Experience taking ML models from experimentation through to deployment
  • Familiarity with robotic teleoperation and data collection techniques
  • A strong experimental mindset and the ability to work effectively in a fast-moving environment

Research contributions to conferences such as ICRA, CoRL, RSS, NeurIPS or CVPR would be valuable, but they're not essential.

Experience with embedded AI or edge deployment, open-source robotics or ML projects, production robotics, industrial automation or tactile sensing would also be advantageous.

The environment

This is an early-stage robotics team where research and engineering sit very close together.

You'll work alongside engineers who are focused on building practical robotic systems, with the freedom to experiment but also the expectation that promising ideas eventually need to work outside of simulation.

As a Founding Robotics ML Research Engineer, you'll have a meaningful voice in the research direction, engineering stack and development culture of the company.

There is no established playbook telling you which model to use, which architecture to follow or how the system should ultimately work. You'll help figure that out.

The team is looking for people who move quickly, think deeply and care about whether their technology actually works.

Interested?

If you're a Founding Robotics ML Research Engineer who wants to work at the frontier of robot learning while seeing your research deployed on physical systems, we'd like to hear from you.

Apply through this advert or contact our team directly for a confidential conversation about the opportunity.