Research Scientist, Engineer – Reinforcement Learning Infrastructure

Luma AI

. Design, build, and scale distributed RL post-training systems across thousands of GPUs .

Posted 10/8/2026full-timeLondon • United KingdomMid-LevelSeniorWebsite

Core Competencies

Role fit

Core Competencies

Use this summary to align your resume positioning with the role.

Demonstrates expertise in designing and scaling distributed Reinforcement Learning (RL) systems, with a strong focus on post-training LLMs and high-throughput rollout generation. Proficient in building RL environments and reward infrastructures while ensuring stability and efficiency across large GPU clusters.

Highest-signal resume keywords

Post-Training LLMs ExperienceDistributed PyTorch TrainingRL Environment DevelopmentGPU Clusters UnderstandingRL Post-Training Frameworks Familiarity

ATS Keywords

Tailor your resume

Applicant Tracking System Keywords

Tip: use these terms in your resume and cover letter to boost ATS matches.

Hard Skills

Reinforcement Learning (PPO/GRPO-family, RLHF, RLVR)Distributed PyTorchWeight SynchronizationReward Infrastructure DevelopmentEvaluation HarnessesMulti-Turn Tool UseAsynchronous ArchitecturesDebugging ToolingRollout Inference EnginesReward Functions

Tools & Technologies

VLLMSGLangNCCLMPIKubernetesRay

Industry Keywords

Large-Scale RLGPU ClustersMixed Training WorkloadsContainerizationOrchestration

Tech Stack

Tools & technologies

KubernetesPyTorchRay

About the role

Key responsibilities & impact

Design, build, and scale distributed RL post-training systems across thousands of GPUs
Build high-throughput rollout generation integrating vLLM, SGLang, weight synchronization, and asynchronous/off-policy schemes
Design RL environments for agentic, multi-step tasks including sandboxed code execution, tool use, computer use, and multimodal interaction
Build reward infrastructure including verifiable/programmatic rewards, reward-model serving, LLM-as-judge pipelines, and defenses against reward hacking
Develop evaluation, monitoring, and debugging tooling for stable large-scale RL runs
Advance training efficiency and stability and turn post-training ideas into production runs with researchers
Learn the current RL stack and diagnose throughput, stability, and correctness issues
Improve and validate a component of the RL loop on a real run
Harden the full loop across thousands of GPUs and asynchronous architectures

Requirements

What you’ll need

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
Nice to have: containerization and orchestration (Kubernetes, Ray) for large environment fleets and sandboxed workloads
Nice to have: research contributions in RL for LLMs or open-source contributions to RL training frameworks

Benefits

Comp & perks

Equal opportunity employer
Voluntary diversity and inclusion survey; participation does not affect the job application