RL Systems Engineer: Scalable Post-Training Infrastructure

Luma; United Kingdom; On-site; GBP 150,000 - 190,000; Full time

Job summary

Luma in the United Kingdom is building RL systems at frontier scale, coupling policy optimization with thousands of GPUs. You will design, build, and scale post-training pipelines that coordinate trainer, rollout, environments, and reward workloads.

We seek someone with hands-on experience training LLMs with RL, expertise in distributed PyTorch, and familiarity with inference engines like vLLM and SGLang. You’ll help harden the loop, improve throughput, and ship robust evaluation tools.

Qualifications

• Hands-on experience with post‑training LLMs using RL (PPO/GRPO family, RLHF, RLVR).
• Extensive distributed PyTorch training and model parallelism for foundation models.
• Experience building RL environments, reward functions, verifiers, or evaluation harnesses for LLM agents.
• Familiarity with RL post‑training frameworks (veRL, OpenRLHF, TRL) and rollout inference engines.
• Strong understanding of GPU clusters, networking, and communication libraries (NCCL, MPI).

Responsibilities

• 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.
• Build reward infrastructure: verifiable/programmatic rewards, reward‑model serving, LLM‑as‑judge pipelines, and defenses against reward hacking.
• 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.
• 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 and prove it on a real run.
• Days 60–90 — Scale & Systemize: Harden the full loop across thousands of GPUs and asynchronous architectures.

Skills

Post-training RL

Distributed PyTorch

RL environments

GPGPU systems

RLHF / PPO

Tools

vLLM

SGLang

veRL

OpenRLHF

TRL

Ray

Kubernetes

NCCL

MPI

PyTorch

CUDA

Job description

Luma in the United Kingdom is building RL systems at frontier scale, coupling policy optimization with thousands of GPUs. You will design, build, and scale post-training pipelines that coordinate trainer, rollout, environments, and reward workloads.

We seek someone with hands-on experience training LLMs with RL, expertise in distributed PyTorch, and familiarity with inference engines like vLLM and SGLang. You’ll help harden the loop, improve throughput, and ship robust evaluation tools.