Research Scientist / Engineer – Reinforcement Learning Infrastructure
Lumaai · London, UK
What this role requires
3 requirements, read out of the advert rather than guessed from the job title:
Also mentioned, not required: Kubernetes, Ray, NCCL, MPI, English. Worth having, but their absence is not what gets a CV filtered out.
See how often each of these is required across open data roles in Europe.
Free, no card. Tells you which of them your CV evidences and which it only implies.
Job description
You'll build the systems that make reinforcement learning work at frontier scale — coupling policy optimization with large fleets of inference workers, agentic environments, and the reward and verification systems that turn model behavior into learning signal. RL is how Luma's models go from capable to useful. RL at scale is a full-loop systems problem: training, rollout generation, environment execution, and reward computation running concurrently across thousands of GPUs, all needing to stay fast, stable, and correct together. It fits someone who has lived this — post-trained LLMs with RL, built environments and verifiers, and debugged asynchronous rollout pipelines at scale. If you haven't operated RL at real scale, this will be deep water.
What You'll Own
Design, build, and scale distributed RL post-training systems, orchestrating trainer, rollout, environment, and reward workloads across thousands of GPUs.
Build high-throughput rollout generation, integrating inference engines (vLLM, SGLang), weight synchronization, and asynchronous/off-policy schemes.
Design RL environments for agentic, multi-step tasks — sandboxed code execution, tool use, computer use, multimodal interaction — reproducible and scalable to millions of episodes.
Build reward infrastructure: verifiable/programmatic rewards, reward-model serving, LLM-as-judge pipelines, and defenses against reward hacking.
Read the full description (19 more sections)Show less
Develop the evaluation, monitoring, and debugging tooling that keeps large RL runs stable.
Advance training efficiency and stability, and turn new post-training ideas into production runs with researchers.
First 90 Days
One way the first 90 could unfold.
Days 1–30 — Immerse & Diagnose: Learn the current RL stack and where throughput, stability, or correctness break.
Days 30–60 — Ship & Validate: Improve a piece of the loop (rollout throughput, reward infra, or an environment) and prove it on a real run.
Days 60–90 — Scale & Systemize: Harden the full loop across thousands of GPUs and asynchronous architectures.
What You Bring
Hands-on experience post-training LLMs with RL (PPO/GRPO-family, RLHF, RLVR) at meaningful scale.
Extensive distributed PyTorch training and parallelism (FSDP, Tensor/Pipeline/Expert Parallel) for foundation models.
Experience building RL environments, reward functions, verifiers, or evaluation harnesses for LLM agents, including sandboxed execution and multi-turn tool use.
Deep familiarity with RL post-training frameworks (veRL, OpenRLHF, TRL, Ray orchestration) and rollout inference engines (vLLM, SGLang).
Strong understanding of GPU clusters, networking, and communication libraries (NCCL, MPI) under mixed training and inference workloads.
Nice to Have
Running RL training across 100+ GPUs, including asynchronous or disaggregated trainer/rollout architectures.
Containerization and orchestration (Kubernetes, Ray) for large environment fleets and sandboxed workloads.
Research contributions in RL for LLMs, or open-source contributions to RL training frameworks.
About Luma: Luma's mission is to build unified general intelligence that can generate, understand, and operate in the physical world. We believe multimodality is critical for intelligence — the next step beyond language models comes from vision. Luma is an equal opportunity employer.
Find more English Speaking Jobs in United Kingdom on Arbeitnow
You will apply. Then you will hear nothing.
And no one will tell you what was wrong. See it before you send: your ATS score, every weak line, and the fix for each.
- 1
Drop in your CV
One PDF, thirty seconds. No card.
- 2
See what is wrong with it
Every weak passage, quoted from your own CV, with the line to replace it.
- 3
Apply where you fit
Every European role ranked against what your CV actually says.
- Free, no card
- Your CV file is deleted after parsing
- Or skip the upload — build your profile by hand
- Refreshed every 6 hours