GPU Performance Engineer - AI Inference & Acceleration

We Love AI Jobs; Los Angeles, Northern (CA, KY); Hybrid; USD 120,000 - 170,000; Full time

Job summary

HeyGen is seeking a Software Engineer focused on GPU performance to accelerate the speed and efficiency of AI video products. You will profile model execution and inference, implement optimizations, and collaborate with researchers and infrastructure teams to bring improvements into production.

The role emphasizes profiling tools, kernel optimization, and measurable reductions in latency and cost while maintaining output quality.

Qualifications

• Experience optimizing GPU-based AI workloads or high-performance computing systems.
• Proficiency in Python and PyTorch; able to profile and optimize GPU kernels.
• Familiarity with GPU hardware details such as memory bandwidth and tensor cores.

Responsibilities

• Profile GPU workloads to identify bottlenecks in model execution and inference.
• Improve latency, throughput, and GPU utilization through batching and scheduling.
• Develop benchmarks and regression tests to track performance.
• Collaborate with researchers and infrastructure engineers to deploy optimizations.
• Measure impact on latency, throughput, and cost.

Skills

Python

PyTorch

GPU profiling

CUDA

Triton

PyTorch Profiler

Tools

NVIDIA Nsight Systems

NVIDIA Nsight Compute

CUDA

Triton

PyTorch Profiler

Job description

HeyGen is seeking a Software Engineer focused on GPU performance to accelerate the speed and efficiency of AI video products. You will profile model execution and inference, implement optimizations, and collaborate with researchers and infrastructure teams to bring improvements into production.

The role emphasizes profiling tools, kernel optimization, and measurable reductions in latency and cost while maintaining output quality.