Research Scientist / Engineer – Reinforcement Learning Infrastructure

lumaai; United States; Remote; USD 180,000 - 260,000; Full time

Job summary

lumaai is seeking a frontier research engineering role focused on systems to scale reinforcement learning at truly large scale. You will work on the full RL loop—training, rollout generation, environment execution, and reward computation—across thousands of GPUs to ensure throughput, stability, and correctness in production settings.

The role spans post-training LLMs with PPO/GRPO-family methods, RLHF, or RLVR, and involves building environments, reward models, verifiers, and evaluation

Qualifications

• Hands-on experience post-training large language models with RL (PPO/GRPO-family, RLHF, RLVR).
• Extensive distributed PyTorch training with FSDP, Tensor Parallel, Pipeline/Expert Parallel.
• Background building RL environments, reward functions, verifiers, and evaluation harnesses for LLM agents.

Responsibilities

• Design and scale distributed RL post-training systems across thousands of GPUs.
• Build high-throughput rollout generation with inference engines and asynchronous schemes.
• Develop RL environments with sandboxed execution and multi-turn tool use.
• Construct reward infrastructure, including LLM-as-judge pipelines and defenses against reward hacking.
• Build evaluation, monitoring, and debugging tooling for long RL runs.
• Improve training efficiency and stability and deploy new ideas into production with researchers.

Skills

Post-training RL with LLMs

Distributed PyTorch training

RL environments and reward functions

RLHF/RLVR/GRPO methods

GPU cluster management

Tools

veRL

OpenRLHF

TRL

Ray orchestration

Job description

Role overview

A frontier research engineering role focused on the systems that make reinforcement learning practical at truly large scale. The work spans the full RL loop — training, rollout generation, environment execution, and reward computation — distributed across thousands of GPUs, where throughput, stability, and correctness must all hold together at the same time. This is deep systems work for someone who has already post-trained LLMs with RL, not a modeling-only or research-only seat.

Responsibilities

• Design and scale distributed RL post-training systems that orchestrate trainer, rollout, environment, and reward workloads across thousands of GPUs.
• Build high-throughput rollout generation integrating inference engines, weight synchronization, and asynchronous or off-policy schemes.
• Develop RL environments for agentic, multi-step tasks, including sandboxed code execution, tool use, computer use, and multimodal interaction, reproducible at millions of episodes.
• Construct reward infrastructure covering verifiable and programmatic rewards, reward-model serving, LLM-as-judge pipelines, and defenses against reward hacking.
• Build evaluation, monitoring, and debugging tooling that keeps long RL runs stable and diagnosable end to end.
• Improve training efficiency and stability, and convert new post-training ideas into production runs alongside research collaborators.

Requirements

• Hands-on experience post-training large language models with RL, covering PPO/GRPO-family methods, RLHF, or RLVR at meaningful scale.
• Extensive distributed PyTorch training with parallelism techniques such as FSDP, Tensor Parallel, Pipeline Parallel, or Expert Parallel for foundation models.
• Background 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 such as veRL, OpenRLHF, TRL, or Ray orchestration, plus rollout inference engines like vLLM or SGLang.
• Strong understanding of GPU clusters, networking, and communication libraries such as NCCL or 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 experience with Kubernetes and Ray for large environment fleets and sandboxed workloads.
• Research contributions in RL for LLMs, or open-source contributions to RL training frameworks.