Research Scientist / Engineer – Performance Optimization

Luma AI

. Profile and optimize GPU, CPU, and accelerator code for maximum utilization and minimal latency .

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 expert-level proficiency in Triton and CUDA programming for GPU optimization, with a strong focus on high-performance PyTorch development and transformer model optimization. Capable of building performance monitoring tools and implementing advanced optimization techniques to enhance model architectures and deployment efficiency.

Highest-signal resume keywords

Expert-Level Triton ProgrammingCUDA OptimizationHigh-Performance PyTorch DevelopmentProfiling Tools ProficiencyTransformer Architecture Understanding

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

Triton ProgrammingCUDA ProgrammingPyTorch DevelopmentKernel DevelopmentCustom OperationsPerformance OptimizationModel Architecture OptimizationDistributed DeploymentProfiling TechniquesAdvanced CUDA Optimization

Tools & Technologies

NVIDIA NsightTorch ProfilerTensorRTONNXXLA

Industry Keywords

GPU OptimizationTransformer ModelsAttention MechanismsKernel Fusion TechniquesInference Workloads

Tech Stack

Tools & technologies

Node.jsPyTorch

About the role

Key responsibilities & impact

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

Requirements

What you’ll need

Expert-level Triton/CUDA programming and GPU optimization
Strong PyTorch skills, including kernel development and custom operations
Proficiency with profiling tools, including NVIDIA Nsight, torch profiler, and custom tooling
Deep understanding of transformer architectures and attention mechanisms
Experience with compilers and exporters such as torch.compile, TensorRT, ONNX, or XLA (nice to have)
Experience optimizing inference workloads for latency and throughput (nice to have)
Knowledge of Triton compiler and kernel fusion techniques (nice to have)
Knowledge of warp-level intrinsics and advanced CUDA optimization (nice to have)

Benefits

Comp & perks

Equal opportunity employer
Hybrid work arrangement
Remote work option in the UK (listed as a secondary location)