Job Title:
Large-Behaviour Models Research Engineer
Location:
Bristol, UK | Hybrid (Primarily On-site)
Team:
Manipulation & Robot Learning Team
Employment Type:
Full-time, Permanent
About the Role
We are pushing the frontier of dexterous, whole-body manipulation. By leveraging large-scale demonstration data — from teleoperation, human demonstrations, and the open web — we enable robots to learn new tasks from just a handful of examples.
As a
Large-Behaviour Models Research Engineer
, you will help design, train, and deploy the next generation of multitask manipulation policies.
What You'll Do
* Invent & apply ML
– Translate cutting-edge behaviour cloning and generative model research into practical solutions for real-world, bimanual manipulation challenges.
* Build great software
– Develop well-documented, tested, and maintainable Python/C++ code that the entire team can rely on.
* Drive technical direction
– Lead design reviews and collaborate closely with Kinisi platform engineers to ensure seamless end-to-end integration.
* Own deployment
– Deploy models onto Kinisi hardware, debug in the loop, and iterate to optimise real-world performance.
Qualifications
* M.S. (or higher) in Computer Science, Machine Learning, Robotics, or a related field.
* Hands-on experience in training and deploying learned control policies — either on physical robots or in complex simulation environments.
* Proficiency with modern machine learning stacks, model architectures, and experimental workflows.
* Strong analytical skills and the ability to diagnose complex runtime issues effectively.