Remote GPU Software Engineer - AI Acceleration

YO IT Consulting; Dubai; Remote; AED 246,000 - 402,000; Full time

Job summary

YO IT Consulting seeks an experienced GPU Software Engineer in Dubai to design and optimize AI acceleration software. You will implement compute solutions with CUDA, WebGPU, and GLSL, develop high-performance kernels, and profile workloads to remove bottlenecks within NVIDIA GPU architectures.

The role focuses on host-device integration, shader and kernel optimization, and scalable AI/LLM workloads in a remote contractor setting. Strong C skills and HPC background are essential.

Qualifications

• Strong hands-on experience with GPU programming on NVIDIA hardware.
• Proficiency in CUDA, WebGPU, GLSL or equivalents.
• Experience optimizing compute kernels and shader code.

Responsibilities

• Design GPU-accelerated software for AI, LLM, graphics, scientific computing, and other high-performance workloads.
• Implement compute solutions using CUDA, WebGPU, GLSL, or related GPU programming technologies.
• Develop and optimize GPU kernels for improved execution speed and resource efficiency.
• Profile computational workloads to identify performance limitations and inefficient execution patterns.
• Analyze GPU utilization, kernel behavior, and memory-related bottlenecks.
• Apply targeted optimization strategies based on profiling results and hardware characteristics.
• Develop host-side software and GPU integration logic using C .
• Build reliable interfaces between CPU-side application code and GPU-accelerated components.
• Optimize data movement and execution flow between host and device environments.
• Create GPU-focused programming tasks and technical solutions for AI and LLM applications.
• Evaluate whether proposed implementations make effective use of available GPU resources.
• Refactor performance-critical code while maintaining correctness and maintainability.
• Develop and optimize shaders for compute or graphics-oriented workloads where applicable.
• Apply understanding of NVIDIA GPU architecture when making optimization decisions.
• Benchmark alternative implementations and compare performance characteristics.
• Identify opportunities to reduce execution overhead and improve computational throughput.
• Investigate performance regressions and determine their underlying technical causes.
• Contribute technical insight to GPU architecture and acceleration discussions.
• Document optimization decisions, performance findings, and implementation trade-offs.
• Collaborate with distributed technical teams on demanding GPU software challenges.

Skills

GPU programming

NVIDIA GPUs

CUDA

WebGPU

GLSL

C

Kernel development

Kernel optimization

Shader programming

Shader optimization

GPU profiling

Performance optimization

Parallel computing

High-performance computing

AI acceleration

LLM acceleration

Machine learning acceleration

Scientific computing

Graphics programming

Host-device integration

GPU memory optimization

Performance benchmarking

Bottleneck analysis

Compute workloads

Code optimization

Performance engineering

GPU software development

Job description

YO IT Consulting seeks an experienced GPU Software Engineer in Dubai to design and optimize AI acceleration software. You will implement compute solutions with CUDA, WebGPU, and GLSL, develop high-performance kernels, and profile workloads to remove bottlenecks within NVIDIA GPU architectures.

The role focuses on host-device integration, shader and kernel optimization, and scalable AI/LLM workloads in a remote contractor setting. Strong C skills and HPC background are essential.