ResearchLondon / RemoteFull-time
Applied Research Scientist
Our research team exists to make the platform smarter, faster and safer — not to chase leaderboard glory. You'll run experiments that ship: routing policies, eval harnesses, fine-tuning recipes and safety systems used by 12,000+ teams within weeks of working.
What you'll do
- Design and run experiments on model routing, distillation and fine-tuning
- Build eval suites that measure what users actually feel
- Publish internally first, externally when it helps the field
- Pair with platform engineers to move results into production
- Advise product teams on capability and safety trade-offs
What we're looking for
- PhD or equivalent research experience in ML, NLP or related fields
- Strong empirical skills: you design clean experiments and distrust noise
- Fluency with PyTorch or JAX and modern training infrastructure
- Publication record or public artifacts demonstrating rigor
- Ability to communicate findings to non-research audiences
Nice to have
- Experience with RLHF, DPO or preference optimization
- Work on multimodal (vision, speech or video) models
- Open-source research tooling contributions
Apply
Apply for Applied Research Scientist
Five fields, five minutes. We review every application by hand and reply within a week — usually faster.