Luma AI

Research Scientist

• Profile and optimize GPU, CPU, and accelerator code for maximum utilization and minimal latency
• Write high-performance PyTorch, Triton, and CUDA, including custom operations when needed
• Develop fused kernels and leverage tensor cores and modern hardware features across platforms
• Optimize model architectures and implementations for distributed multi-node production deployment
• Build performance monitoring and analysis tools and automation
• Research and implement cutting-edge optimization techniques for transformer models
• Profile current training and inference paths during the first 30 days
• Ship and validate a kernel or architecture optimization during days 30–60
• Build monitoring and automation to prevent performance regressions during days 60–90

Explore Research Scientist – Performance Optimization at Luma AI. Review requirements and application details.

Research Scientist – Performance Optimization

• 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 scalable 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, diagnose bottlenecks, ship and validate improvements, and harden the full loop across thousands of GPUs

Explore Research Scientist / Engineer – Reinforcement Learning… at Luma AI. Review requirements and application details.

Research Scientist / Engineer – Reinforcement Learning Infrastructure

• Design, implement, and optimize efficient distributed training systems for models across thousands of GPUs
• Research and implement advanced parallelization, including FSDP, Tensor Parallel, Pipeline Parallel, and Expert Parallel
• Build monitoring, visualization, and debugging tools for large-scale training runs
• Optimize training stability, convergence, and resource utilization across massive clusters
• Learn the current training stack and diagnose stability and utilization issues at scale
• Deliver a parallelization or stability improvement that measurably benefits a real training run
• Build monitoring and tooling to keep large runs reliable and efficient

Explore Research Scientist / Engineer – Training Infrastructure at Luma AI. Review requirements and application details.

Research Scientist / Engineer – Training Infrastructure

• Design, implement, and optimize efficient distributed training systems for models across thousands of GPUs
• Research and implement advanced parallelization, including FSDP, Tensor Parallel, Pipeline Parallel, and Expert Parallel
• Build monitoring, visualization, and debugging tools for large-scale training runs
• Optimize training stability, convergence, and resource utilization across massive clusters
• Learn the current training stack and diagnose stability and utilization issues at scale during the first 30 days
• Ship and validate a parallelization or stability improvement that measurably helps a real training run during days 30–60
• Build monitoring and tooling to keep large runs reliable and efficient during days 60–90
• Build distributed systems that train Luma's large-scale multimodal models across thousands of GPUs
• Enable researchers to focus on innovation on top of reliable, efficient, scalable infrastructure

• Profile and optimize GPU, CPU, and accelerator code for maximum utilization and minimal latency
• Write high-performance PyTorch, Triton, and CUDA code, including custom operations when needed
• Develop fused kernels and leverage tensor cores and modern hardware features across platforms
• Optimize model architectures and implementations for distributed multi-node production deployment
• Build performance monitoring and analysis tools and automation
• Research and implement cutting-edge optimization techniques for transformer models
• Profile current training and inference paths during the first 30 days
• Deliver a kernel or architecture optimization that measurably improves utilization or latency during days 30–60
• Build monitoring and automation to prevent performance regressions during days 60–90

Explore Research Scientist / Engineer – Performance Optimization at Luma AI. Review requirements and application details.

Research Scientist / Engineer – Performance Optimization

• 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

Explore Research Scientist, Engineer – Reinforcement Learning… at Luma AI. Review requirements and application details.

Research Scientist, Engineer – Reinforcement Learning Infrastructure

Full-stack Engineer

• Own how Luma's models get served by integrating new architectures into the inference engine
• Scale deployments across thousands of machines
• Keep expensive GPU fleets utilized while meeting internal SLOs
• Collaborate across research, engineering, and infrastructure to optimize model efficiency and deployments
• Build internal tooling to measure, profile, and track inference jobs and workflows
• Automate, test, and maintain inference services for uptime and reliability
• Manage and optimize inference workloads across clusters and hardware providers
• Build scheduling systems that optimize GPU resources while meeting SLOs
• Maintain CI/CD for model checkpoints and SDKs
• Learn the inference stack, fleets, and reliability or utilization issues during the first 30 days
• Integrate a model or ship tooling/scheduling improvements during days 30–60
• Harden deployment pipelines and scheduling across clusters and providers during days 60–90

Explore Software Engineer, Inference at Luma AI. Review requirements and application details.

Software Engineer, Inference

• Own how Luma's models get served through inference-engine integration, deployment scaling, and GPU fleet utilization
• Integrate new model architectures into the inference engine
• Collaborate across research, engineering, and infrastructure to optimize model efficiency and deployments
• Build internal tooling to measure, profile, and track inference jobs and workflows
• Automate, test, and maintain inference services for uptime and reliability
• Manage and optimize inference workloads across clusters and hardware providers
• Scale deployments across thousands of machines
• Build scheduling systems that optimize GPU resources while meeting SLOs
• Maintain CI/CD for model checkpoints and SDKs
• Learn the inference stack and diagnose reliability or utilization issues during the first 30 days
• Integrate a model or ship tooling/scheduling improvements during days 30–60
• Harden deployment pipelines and scheduling across clusters and providers during days 60–90