Join Us
Join the research
We hire researchers with serious technical depth and engineers who understand that the infrastructure they build shapes the science. We are small by design and intend to stay that way.
Open positions
Research Scientist — Self-Supervised Learning
Full-time · London, UK
We are looking for a researcher with deep experience in self-supervised learning. You will work on contrastive and non-contrastive objectives, masked prediction, training stability, and the geometric properties of unsupervised embedding spaces. You will also contribute to our evaluation methodology for representations formed without labelled data. Prior publications in self-supervised or unsupervised representation learning are expected.
Research Engineer — Training Infrastructure
Full-time · London, UK
We need an engineer who can build and maintain distributed training systems for large unsupervised models. Experience with PyTorch distributed training, mixed-precision training, gradient checkpointing, and compute cluster management required. We are not looking for someone who only implements; research engineering here means making design decisions that affect what experiments are feasible.
Research Intern (Year-round)
Internship · London, UK · 3–6 months
We welcome PhD students and final-year undergraduates with research interests in self-supervised learning, reward-free reinforcement learning, or continual learning. Interns are assigned a defined project within one of our active research threads and are treated as full members of the group. We expect written output at the end of a placement, whether or not results are positive.
Apply via email
Send your CV, a brief covering note describing your research interests and why you want to work on these problems, and links to any relevant papers or code. We do not use application portals. Direct email is the right channel.
[email protected]What it is like here
We are a small group and we intend to remain one. Every researcher owns a strand of the programme and is expected to define their own questions within it. There is no layer of management between researchers and research decisions. We discuss, we disagree, we test ideas empirically. This is not a place to execute a predetermined roadmap.
We work in London. We think proximity matters for foundational research — the informal conversations that move ideas forward do not happen well at a distance. We are not inflexible about specific days, but we are clear that this is a London position.
We are independently funded. This means competitive salaries, no equity component (we have nothing to vest and no intention of creating that kind of pressure), a generous compute allocation, and no product deadline on the science. The constraint is that we are a small organisation: there is no large HR department, no structured performance review process, and no path to promotion through a bureaucratic ladder. If those things matter to you, a larger institution may be a better fit.