AI Researcher — Applied Frontier AI Lab, London
A new applied frontier AI lab in London, founded by a team from a major AI research lab. Around 15 people today, growing to 35–40 over the next 18 months.
This isn't a pure research lab and it isn't a product company — it's both. The work has to be original, and it has to ship. We're looking for people who can do research at the frontier and build it into something real, not hand it off.
We're interested in three research pillars:
- Memory & long-context systems — retrieval, long-context reasoning, episodic memory, model-based RL
- Alignment & self-improvement — reward modelling, preference learning, agentic RL, on-policy training
- Interpretability & oversight — mechanistic interpretability, sparse autoencoders, representation engineering, scalable oversight
What we're looking for
Research scientists and research engineers from frontier labs (DeepMind, Anthropic, OpenAI, FAIR) or fast-moving applied AI startups, at any seniority. Strong technical depth in one of the above areas. JAX experience is a meaningful signal.
Beyond the CV, we care about:
- High agency — you chose the problems you worked on, you weren't just assigned them
- Fast movement — you've moved quickly through your career and left environments that were too slow for you
- Product instinct — you think about what your research does in the real world, not just whether it's interesting
The opportunity
Join as a founding researcher, not employee #100. Set research direction, shape the next twenty hires, own a pillar.
Competitive base + meaningful equity. Sponsorship available, no nationality restrictions. London-based, five days in office.
Interested? Get in touch in confidence via Avanti Recruitment.
