Machine Learning Research Scientist

Machine Learning Research Scientist

Can we build robots that learn and generalise, rather than simply execute what they’ve been trained to do?

We’re working with an early-stage UK robotics company building learning systems designed to operate in complex, unpredictable real-world environments.

They’re hiring a Machine Learning Research Scientist to tackle fundamental problems in robot intelligence - with research spanning vision-language-action models, world models, representation learning and robot learning.

The questions are genuinely open-ended.

How should robots represent and reason about the physical world? How can they learn new behaviours efficiently? And how do we build models that generalise beyond the environments and tasks they were trained on?

Your research could involve:

• Developing new VLA and world model architectures
• Exploring representations for embodied intelligence
• Improving generalisation and transfer across tasks and environments
• Designing and running large-scale learning experiments
• Taking promising ideas from papers and first principles into new research directions
• Testing models on real robotic systems, not just offline benchmarks

You’ll join early enough to have genuine influence over the research agenda, technical direction and systems being built, alongside meaningful early-stage equity.