Reinforcement Learning Engineer at lumalabs-ai

Location: Colorado City, Arizona
Level: not_specified
Posted: Jul 21, 2026

Description

Luma AI is seeking experienced Reinforcement Learning Engineers to join their RL Infrastructure team. This role involves designing, building, and scaling distributed RL post-training systems for large multimodal models, optimizing high-throughput rollout generation, implementing RL environments for complex tasks, developing reward infrastructure, and creating evaluation tools. Candidates should have hands-on experience with RL techniques like PPO/GRPO-family methods, RLHF, and RLVR, extensive knowledge of distributed PyTorch training, and deep familiarity with RL post-training frameworks.

Required Skills

Reinforcement Learning
PyTorch Distributed Training
FSDP
Tensor/Pipeline/Expert Parallel
veRL
OpenRLHF
TRL
Ray-based Orchestration

Additional Information

Required Previous Experiences: Experience post-training LLMs with reinforcement learning (e.g. PPO / GRPO-family methods, RLHF, RLVR / RL from verifiable rewards)
Extensive experience with distributed PyTorch training and parallelization strategies (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 and their systems tradeoffs (e.g. veRL, OpenRLHF, TRL, Ray-based orchestration) and inference engines used for rollouts (vLLM, SGLang)
Strong understanding of GPU clusters, networking, and communication libraries (NCCL, MPI), and how they behave under mixed training + inference workloads
Preferred Qualifications: Experience running RL training across >100 GPUs, including asynchronous or disaggregated trainer/rollout architectures
Experience with containerization and orchestration (Kubernetes, Ray) for large environment fleets and sandboxed workloads
Research contributions in RL for LLMs — reasoning, agents, reward modeling, or long-horizon tasks — or open-source contributions to RL training frameworks